A method for automatically constructing a three-dimensional model for intelligently identifying active faults

Abstract

The present invention discloses a method for automatically constructing a three-dimensional model for intelligently identifying active faults, which belongs to the field of three-dimensional modeling. The nearest neighbor index method is used to determine whether a small earthquake catalog meets the small earthquake clustering property; on the basis of meeting the small earthquake clustering property, the critical distance and noise threshold of the hierarchical clustering are set to obtain a clustered small earthquake catalog; along the direction of the clustered earthquake, a high-density three-dimensional slice with a custom step size is established and the overlap of each profile projection is set to obtain seismic data for characterizing the fault, and random sampling is used to consistently fit the earthquake to obtain the fault interpretation line; a three-dimensional initial model of the active fault is constructed by a spatial interpolation method; the spatial relationship of each branch fault is determined to construct the final three-dimensional model of the active fault. The present invention adopts the above-mentioned method for automatically constructing a three-dimensional model for intelligently identifying active faults, and can directly construct a three-dimensional model of the active fault based on the small earthquake catalog, providing a new method for characterizing the three-dimensional model of the active fault.

Description

Technical Field

[0001] The present invention relates to the technical field of three-dimensional modeling, and in particular to a method for automatically constructing a three-dimensional model for intelligently identifying active faults. Background Art

[0002] Precise positioning of earthquake sequences is an important means to determine the deep and shallow geometric morphology of earthquake-causing faults. According to the law that clusters of small earthquakes occur on and near the fault plane of a large earthquake, a three-dimensional fault model can be established using the location of small earthquakes. The key to constructing a three-dimensional model of an active fault based on an earthquake catalog is to extract the small earthquake groups related to the modeled fault and reasonably fit the fine geometric structure of the fault. Traditional methods for characterizing active fault models based on small earthquake catalogs mostly rely on qualitative constraints of small earthquake data and have great subjectivity. The present invention, based on an improved hierarchical clustering method and quantitative judgment of the clustering of small earthquake catalogs, obtains a new quantitative method to obtain a high-quality small earthquake catalog. The fault interpretation line is fitted by high-density three-dimensional slicing technology, which overcomes the subjectivity of manual interpretation and obtains a quantitative two-dimensional fault structure. This method is a brand-new workflow that can directly construct a three-dimensional model of an active fault based on a small earthquake catalog, providing a new method for characterizing a three-dimensional model of an active fault. Summary of the invention

[0003] The purpose of the present invention is to provide a method for automatically constructing a three-dimensional model for intelligently identifying active faults, so as to solve the problems mentioned in the above background technology.

[0004] To achieve the above object, the present invention provides a method for automatically constructing a three-dimensional model for intelligently identifying active faults, comprising the following steps:

[0005] S1. Obtain data on the epicenter area, including surface fault traces, earthquake catalogs, and focal mechanism solutions of related earthquakes; analyze the minimum integrity magnitude based on the earthquake relocation data in the data, extract the complete earthquake sub-catalog, and quantitatively determine whether the small earthquake catalog meets the small earthquake clustering characteristics through the nearest neighbor index method;

[0006] S2. On the basis that the small earthquake catalog conforms to the clustering of small earthquakes, the clustered small earthquake catalog is obtained by setting the critical distance and noise threshold of hierarchical clustering;

[0007] S3. Fitting the optimal fault interpretation line. By establishing high-density three-dimensional slices with a custom step size and setting the overlap of each section projection along the direction of the cluster earthquake, we can obtain rich seismic data describing the fault. Then, we use random sampling to consistently fit the fault line to obtain an objective two-dimensional geometric structure of the fault.

[0008] S4, constructing a three-dimensional initial model of the active fault by spatial interpolation method according to the two-dimensional geometric structure of the fault obtained in S3;

[0009] S5. For a fault zone containing multiple branch faults, repeat S4, combine geological information, determine the spatial relationship of each branch fault, and construct the final three-dimensional fine model of the active fault.

[0010] Preferably, S1 specifically includes:

[0011] S11. Obtain data such as surface fault traces, earthquake catalogs, and focal mechanism solutions of related earthquakes in the epicenter area;

[0012] S12. Using the maximum curvature method, analyze the minimum complete magnitude of the acquired earthquake catalog, extract its minimum complete subcatalog, and obtain a small earthquake catalog;

[0013] S13. Determine whether the small earthquake catalog is clustered.

[0014] Preferably, S13 specifically includes:

[0015] S131, calculate the initial critical distance d from each earthquake event to its nearest neighbor earthquake event min ;

[0016] S132, for all initial critical distances d min According to the number of earthquake events n in the model, the average nearest neighbor distance d is calculated min :

[0017]

[0018] Among them, s i is the earthquake event in the study area, n is the number of earthquake events;

[0019] S133. Under completely random distribution, the theoretical average distance E(d m in) is:

[0020]

[0021] Where V is the volume of the study area and n is the number of earthquake events;

[0022] S134, calculate the nearest neighbor index NNI, NNI is the average nearest neighbor distance d min The expected average nearest neighbor distance E(d min ) ratio:

[0023]

[0024] S135. Determine the distribution pattern of the earthquake catalog based on the value of NNI:

[0025] d. NNI < 1, indicating that the earthquake event is spatially clustered;

[0026] e, NNI = 1, indicating that the earthquake event is randomly distributed;

[0027] NNI>1 indicates that earthquake events tend to be evenly distributed.

[0028] Preferably, according to the principle that clustered small earthquakes occur on and near the fault plane of a large earthquake, the fault rupture surface is assumed to be a curved surface, and the small earthquakes are distributed in a destruction zone and are approximately symmetrically distributed with the fault surface. S2 specifically includes:

[0029] S21. Set the critical distance list d_list: set the initial critical distance d min is 1000m, the step length is 100 meters, and its 5 times value is 5*d min As the maximum critical distance d max , forming the critical distance list d_list;

[0030] S22. Set the noise threshold list ZY_list: Combined with the clustering of small earthquake complete sub-directories, set the minimum noise threshold ZY min , the noise threshold step is ZY_step, and the maximum noise threshold is ZY max , forming a noise threshold list ZY_list;

[0031] S23, select the initial critical distance d in the critical distance list d_list min As the distance threshold for stopping clustering of agglomerative hierarchical clustering, the initial clustering result of clustering is obtained;

[0032] S24, select the minimum noise value ZY in the noise threshold list ZY_list min As the noise threshold for screening earthquake clusters, in the initial clustering results obtained by S23, clusters with the number of earthquake events in each cluster greater than the noise threshold are identified as earthquake clusters; clusters with the number of earthquake events less than or equal to the noise threshold are identified as noise events, and a temporary clustering result is obtained. Then, each noise threshold in the noise threshold list is traversed to obtain the corresponding temporary clustering result.

[0033] S25, selecting the next critical distance value from the critical distance list d_list in turn as the distance threshold for stopping clustering of the hierarchical clustering, obtaining the initial clustering result of the clustering, using the selected critical distance value, executing S24, and obtaining a set of temporary clustering results;

[0034] S26, repeat the above process, traverse all critical distance values ​​in the critical distance list, execute S24 for each critical distance value, and obtain a corresponding temporary clustering result until all critical distance values ​​in the list are traversed;

[0035] S27, based on all temporary clustering results obtained in S26, combined with actual geological information, select the critical distance and noise threshold corresponding to the clustering result with the largest number of earthquake clusters when the earthquake clusters are consistent in strike and dip, which are the optimal critical distance d and optimal noise threshold ZY_n of hierarchical clustering;

[0036] S28. Obtain the final small earthquake catalog clustering result based on S27.

[0037] Preferably, S3 specifically includes:

[0038] S31, according to the clustering result of the small earthquake catalog obtained in S28, the identified trend of each small earthquake cluster is used to set the rotation angle dip of the three-dimensional slice and the interval step of each slice to establish a high-density three-dimensional slice;

[0039] S32, custom setting each slice section to query the earthquake overlap TY_overlap, where the overlap TY_overlap is set as a percentage of the interval step;

[0040] S33. Project the small earthquakes onto each slice profile, and fit the two-dimensional interpretation line of the fault on each profile consistently through random sampling, wherein the number of samples consistent with random sampling is set to 90% of the number of earthquakes on the participating profiles, the number of iterations is 1000, and the model with the smallest error is selected as the optimal fault line.

[0041] Preferably, S4 specifically includes:

[0042] S41. First, determine the nature of the active fault (including exposed surface faults or concealed active faults); based on the two-dimensional geometric structure of the active fault obtained in S3, combined with the surface fault traces in the study area, the focal mechanism solutions of related earthquakes, and the surface rupture conditions obtained, determine whether the fault is an exposed surface fault;

[0043] S42, constructing a three-dimensional initial model of the active fault; using a discrete element interpolation algorithm or other spatial interpolation method to smooth the multiple fault interpretation lines obtained in S3, and generating a three-dimensional initial model of the active fault.

[0044] Preferably, the interpolation algorithm and smoothing processing of the three-dimensional fault plane in S42 meet the following requirements:

[0045] d. Use interpolation algorithms that meet geological and mathematical constraints;

[0046] e. Generate triangular mesh or quadrilateral mesh nodes;

[0047] f. Smoothing is constrained by boundary nodes.

[0048] Preferably, S5 specifically includes:

[0049] S51, for a fault zone including multiple branch faults, repeat S4 to obtain a three-dimensional initial model of each fault;

[0050] S52. Calculate the three-dimensional attribute model of the active fault, including fault scale and occurrence information.

[0051] Therefore, the present invention adopts the above-mentioned method for automatically constructing a three-dimensional model for intelligently identifying active faults, which has the following beneficial effects:

[0052] (1) It can quantitatively reveal the spatial structure of active faults, analyze the three-dimensional geometric structure, segmentation characteristics and heterogeneity of underground faults, and provide important support for the strong strain analysis of active faults and the assessment of earthquake geological hazards;

[0053] (2) It is not only applicable to a single fault, but also to the 3D modeling of faults with complex intersection relationships;

[0054] (3) A new method for determining small earthquake clusters is proposed, which can directly and quantitatively determine whether small earthquake data have cluster characteristics in three-dimensional space;

[0055] (4) A new hierarchical clustering parameter selection method is proposed to quantitatively identify small earthquake clusters on the modeled fault;

[0056] (5) A new fault plane fitting method is proposed, which can use random sampling to consistently fit the three-dimensional fine structure of active faults.

[0057] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 is an overall flow chart of an embodiment of the present invention;

[0059] Figure 2 A schematic diagram of the spatial distribution of small earthquake relocation data according to an embodiment of the present invention;

[0060] Figure 3 This is a schematic diagram of an improved hierarchical clustering analysis according to an embodiment of the present invention;

[0061] Figure 4 Schematic diagram of the distribution of small earthquake clusters and the fitted fault lines identified by an embodiment of the present invention; (a) is a high-density three-dimensional slice and projection, and (b) is a fault interpretation line based on small earthquake fitting;

[0062] Figure 5 This is a three-dimensional fine model of the active fault according to an embodiment of the present invention. DETAILED DESCRIPTION

[0063] Example

[0064] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0065] Reference Figure 1-Figure 5 The present invention discloses a method for automatically constructing a three-dimensional model for intelligently identifying active faults, comprising the following steps:

[0066] S1. Obtain data on the epicentral area, including surface fault traces, earthquake catalogs, and focal mechanism solutions of related earthquakes; analyze the minimum integrity magnitude based on the earthquake relocation data in the data, extract the complete earthquake sub-catalog, and quantitatively determine whether the small earthquake catalog meets the small earthquake clustering characteristics through the nearest neighbor index method.

[0067] S11. Obtain data such as surface fault traces, earthquake catalogs, and focal mechanism solutions of related earthquakes in the epicenter area;

[0068] S12. Using the maximum curvature method, analyze the minimum complete magnitude of the acquired earthquake catalog, extract its minimum complete subcatalog, and obtain a small earthquake catalog;

[0069] S13. Determine whether the small earthquake catalog is clustered.

[0070] S131, calculate the initial critical distance d from each earthquake event to its nearest neighbor earthquake event min ;

[0071] S132, for all initial critical distances d min According to the number of earthquake events n in the model, the average nearest neighbor distance d is calculated min :

[0072]

[0073] Among them, s i is the earthquake event in the study area, n is the number of earthquake events;

[0074] S133. Under completely random distribution, the theoretical average distance E(d min )for:

[0075]

[0076] Where V is the volume of the study area and n is the number of earthquake events;

[0077] S134. Calculate the nearest neighbor index (NNI), where NNI is the average nearest neighbor distance d min The expected average nearest neighbor distance E(d min ) ratio:

[0078]

[0079] S135. Determine the distribution pattern of the earthquake catalog based on the value of NNI:

[0080] f, NNI < 1, indicating that the earthquake event is spatially clustered;

[0081] g, NNI = 1, indicating that the earthquake event is randomly distributed;

[0082] NNI>1 indicates that earthquake events tend to be evenly distributed.

[0083] S2. On the basis that the small earthquake catalog conforms to the clustering characteristics of small earthquakes, a clustered small earthquake catalog is obtained by setting the critical distance and noise threshold of hierarchical clustering.

[0084] According to the principle that clustered small earthquakes occur on and near the fault plane of a large earthquake, the fault rupture surface is assumed to be a curved surface, and the small earthquakes are distributed in a destruction zone and are approximately symmetrically distributed with the fault surface. S2 specifically includes:

[0085] S21. Set critical distance list (d_list): Set the initial critical distance d min is 1000m, the step length is 100 meters, and its 5 times value (5*d min ) as the maximum critical distance d max , thus forming the critical distance list d_list;

[0086] S22. Set the noise threshold list (ZY_list): Combined with the clustering of the small earthquake complete sub-directory, set the minimum noise threshold ZY min , the noise threshold step is ZY_step, and the maximum noise threshold is ZY max , thus forming a noise threshold list ZY_list. (For simplicity, the noise threshold can be directly set to 1);

[0087] S23, select the initial critical distance (d_list) in the critical distance list (d_list) min ) is used as the distance threshold for stopping clustering of agglomerative hierarchical clustering to obtain the initial clustering result of clustering;

[0088] S24, select the minimum noise value (ZY min) is used as the noise threshold for screening earthquake clusters. In the initial clustering results obtained in S23, clusters with the number of earthquake events in each cluster greater than the noise threshold are identified as earthquake clusters; clusters with the number of earthquake events less than or equal to the noise threshold are identified as noise events. In this way, a temporary clustering result is obtained. Next, each noise threshold in the noise threshold list is traversed to obtain the corresponding temporary clustering result.

[0089] S25, select the next critical distance value from the critical distance list (d_list) in turn as the distance threshold for stopping clustering of the hierarchical clustering, and obtain the initial clustering result of the clustering. Use this critical distance to execute S24 to obtain a set of temporary clustering results.

[0090] S26, repeat the above process, traverse all critical distance values ​​in the critical distance list, execute S24 for each critical distance value, and obtain a corresponding temporary clustering result until all critical distance values ​​in the list are traversed;

[0091] S27. Based on all temporary clustering results obtained in S26 and in combination with actual geological information, the critical distance and noise threshold corresponding to the clustering result with the largest number of earthquake clusters when the earthquake clusters are consistent in strike and dip are selected, which are the optimal critical distance (d) and optimal noise threshold (ZY_n) of hierarchical clustering.

[0092] S28. Obtain the final clustering result of the small earthquake catalog based on S27.

[0093] S3. Fit the optimal fault interpretation line. By establishing high-density three-dimensional slices with a custom step size along the direction of the cluster earthquake and setting the overlap of each section projection, we can obtain rich seismic data that characterize the fault. Then, we use random sampling to consistently fit the fault line to obtain the objective two-dimensional geometric structure of the fault.

[0094] S31, according to S28, the clustering result of the small earthquake catalog is obtained, and the trend of each small earthquake cluster is identified to set the rotation angle dip of the three-dimensional slice and the interval step of each slice, which can be set freely. Here, the interval step is set to 2km to establish a high-density three-dimensional slice;

[0095] S32. Customize the settings of each slice profile to query the earthquake overlap (TY_overlap). Here, the overlap is set to 10%, 20%, etc. with an interval step of 2 km.

[0096] S33. Project the small earthquakes onto each slice profile, and fit the two-dimensional interpretation line of the fault on each profile consistently through random sampling, wherein the number of samples consistent with random sampling is set to 90% of the number of earthquakes on the participating profiles, the number of iterations is 1000, and the model with the smallest error is selected as the optimal fault line.

[0097] S4. Based on the two-dimensional geometric structure of the fault obtained in S3, a three-dimensional initial model of the active fault is constructed by spatial interpolation method.

[0098] S41. First, determine the nature of the active fault; based on the two-dimensional geometric structure of the active fault obtained in S3, combined with the surface fault traces in the study area, the focal mechanism solutions of related earthquakes, and the surface rupture conditions obtained, determine whether it is an exposed surface fault;

[0099] S42, construct a three-dimensional initial model of the active fault; use a spatial interpolation method such as a discrete element interpolation algorithm to smooth the multiple fault interpretation lines obtained in S3 to generate a three-dimensional initial model of the active fault. The interpolation algorithm and smoothing of the three-dimensional fault plane meet the following requirements:

[0100] d. Use interpolation algorithms that meet geological and mathematical constraints;

[0101] e. Generate triangular mesh or quadrilateral mesh nodes;

[0102] f. Smoothing is constrained by boundary nodes.

[0103] S5. For a fault zone containing multiple branch faults, repeat S4, combine geological information, determine the spatial relationship of each branch fault, and construct the final three-dimensional model of the active fault.

[0104] S51, for a fault zone including multiple branch faults, repeat S4 to obtain a three-dimensional initial model of each fault;

[0105] S52. Calculate the three-dimensional attribute model of the active fault, including fault scale (length, cutting depth, area) and occurrence information (strike, dip, inclination).

[0106] Therefore, the present invention adopts the above-mentioned method for automatically constructing a three-dimensional model for intelligently identifying active faults, which can quantitatively reveal the spatial structure of active faults, analyze the three-dimensional geometric structure, segmented characteristics and heterogeneity of underground faults, and provide important support for research such as strong strain analysis of active faults and earthquake geological disaster assessment. This method is not only applicable to single faults, but also to three-dimensional modeling of faults with complex intersection relationships. A new small earthquake cluster determination is proposed, which can directly and quantitatively determine whether small earthquake data has cluster characteristics in three-dimensional space. A new hierarchical clustering parameter selection method is proposed, which can quantitatively identify small earthquake groups of modeled faults. A new fault plane fitting method is proposed, which can use random sampling to consistently fit the three-dimensional fine structure of active faults.

[0107] 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 method for automatically constructing a three-dimensional model for intelligently identifying active faults, characterized in that: The following steps are involved: S1. Obtain data from the epicenter area, analyze the minimum integrity magnitude based on the earthquake relocation data in the data, extract the complete earthquake sub-catalog, and use the nearest neighbor index method to quantitatively determine whether the small earthquake catalog conforms to the small earthquake clustering property; the data includes surface fault traces, earthquake catalogs, and focal mechanism solutions of related earthquakes; S2. On the basis that the small earthquake catalog conforms to the clustering of small earthquakes, the clustered small earthquake catalog is obtained by setting the critical distance and noise threshold of hierarchical clustering; S3, fitting the optimal fault interpretation line, by establishing high-density three-dimensional slices with a custom step size along the direction of the cluster earthquake and setting the overlap of each section projection, to obtain the seismic data describing the fault, and then using random sampling to consistently fit the fault line to obtain the two-dimensional geometric structure of the fault; S4, constructing a three-dimensional initial model of the active fault by spatial interpolation method according to the two-dimensional geometric structure of the fault obtained in S3; S5. For a fault with multiple branch faults, repeat S4, combine geological information, determine the spatial relationship of each branch fault, and construct the final three-dimensional model of the active fault.

2. The method for automatically constructing a three-dimensional model for intelligently identifying active faults according to claim 1, characterized in that: S1 specifically includes: S11. Obtain surface fault traces, earthquake catalogs and focal mechanism solutions of related earthquakes in the epicenter area; S12. Using the maximum curvature method, analyze the minimum complete magnitude of the acquired earthquake catalog, extract its minimum complete subcatalog, and obtain a small earthquake catalog; S13. Determine whether the small earthquake catalog is clustered.

3. The method for automatically constructing a three-dimensional model for intelligently identifying active faults according to claim 2, characterized in that: S13 specifically includes: S131, calculate the initial critical distance d from each earthquake event to its nearest neighbor earthquake event min ; S132, for all initial critical distances d min According to the number of earthquake events in the model, n, the average nearest neighbor distance is calculated. Among them, s i is the earthquake event in the study area, n is the number of earthquake events; S133. Under completely random distribution, calculate the theoretical average distance E(d min ): Where V is the volume of the study area and n is the number of earthquake events; S134, calculate the nearest neighbor index NNI, NNI is the average nearest neighbor distance The expected average nearest neighbor distance E(d min ) ratio: S135. Determine the distribution pattern of the earthquake catalog based on the value of NNI: a. NNI < 1, indicating that the earthquake event is spatially clustered; b, NNI = 1, indicating that earthquake events are randomly distributed; c. NNI>1, indicating that earthquake events tend to be evenly distributed.

4. The method for automatically constructing a three-dimensional model for intelligently identifying active faults according to claim 1, characterized in that: According to the principle that clustered small earthquakes occur on and near the fault plane of a large earthquake, the fault rupture surface is set as a curved surface. The small earthquakes are distributed in a destruction zone and are symmetrically distributed with the fault surface. S2 specifically includes: S21. Set the critical distance list d_list: set the initial critical distance d min is 1000m, the step length is 100 meters, and its 5 times value is 5*d min As the maximum critical distance d max , forming the critical distance list d_list; S22. Set the noise threshold list ZY_list: Combined with the clustering of small earthquake complete sub-directories, set the minimum noise threshold ZY min , the noise threshold step is ZY_step, and the maximum noise threshold is ZY max , forming a noise threshold list ZY_list; S23, select the initial critical distance d in the critical distance list d_list min As the distance threshold for stopping clustering in agglomerative hierarchical clustering, the initial clustering result is obtained; S24, select the minimum noise value ZY in the noise threshold list ZY_list min As the noise threshold for screening earthquake clusters, in the initial clustering results obtained in S23, clusters with the number of earthquake events in each cluster greater than the noise threshold are identified as earthquake clusters; clusters with the number of earthquake events not greater than the noise threshold are identified as noise events, and a temporary clustering result is obtained. Then, each noise threshold in the noise threshold list is traversed to obtain the corresponding temporary clustering result. S25, selecting the next critical distance value from the critical distance list d_list in turn as the distance threshold for stopping clustering of the hierarchical clustering, obtaining an initial clustering result, and using the selected critical distance value to execute S24 to obtain a set of temporary clustering results; S26, repeat the above process, traverse all critical distance values ​​in the critical distance list, execute S24 for each critical distance value, and obtain a corresponding temporary clustering result until all critical distance values ​​in the list are traversed; S27, based on all temporary clustering results obtained in S26, combined with actual geological information, select the critical distance and noise threshold corresponding to the clustering result with the largest number of earthquake clusters when the earthquake clusters are consistent in strike and dip, which are the optimal critical distance d and optimal noise threshold ZY_n of hierarchical clustering; S28. Obtain the final small earthquake catalog clustering result based on S27.

5. The method for automatically constructing a three-dimensional model for intelligently identifying active faults according to claim 1, characterized in that: S3 specifically includes: S31, according to the clustering result of the small earthquake catalog obtained in S28, identify the trend of each small earthquake cluster, set the rotation angle dip of the three-dimensional slice and the interval step of each slice, and establish a high-density three-dimensional slice; S32, custom setting each slice section to query the earthquake overlap TY_overlap, where the overlap TY_overlap is set as a percentage of the interval step; S33. Project the small earthquakes onto each slice profile, and fit the two-dimensional interpretation line of the fault on each profile consistently through random sampling, wherein the number of samples consistent with random sampling is set to 90% of the number of earthquakes on the participating profiles, the number of iterations is 1000, and the model with the smallest error is selected as the optimal fault line.

6. The method for automatically constructing a three-dimensional model for intelligently identifying active faults according to claim 1, characterized in that: S4 specifically includes: S41. Determine the nature of the active fault; based on the two-dimensional geometric structure of the active fault obtained in S3, combined with the surface fault traces in the study area, the focal mechanism solutions of related earthquakes, and the surface rupture conditions obtained, determine whether the fault is an exposed surface fault; S42, constructing a three-dimensional initial model of the active fault; using a discrete element interpolation algorithm, smoothing the multiple fault interpretation lines obtained in S3 to generate a three-dimensional initial model of the active fault.

7. The method for automatically constructing a three-dimensional model for intelligently identifying active faults according to claim 6, characterized in that: The interpolation algorithm and smoothing of the 3D fault plane in S42 meet the following requirements: d. Use interpolation algorithms that meet geological and mathematical constraints; e. Generate triangular mesh or quadrilateral mesh nodes; f. Smoothing is constrained by boundary nodes.

8. The method for automatically constructing a three-dimensional model for intelligently identifying active faults according to claim 1, characterized in that: S5 specifically includes: S51, for a fault zone including multiple branch faults, repeat S4 to obtain a three-dimensional initial model of each fault; S52. Calculate the three-dimensional attribute model of the active fault, including fault scale and occurrence information.

Images