A water-rich fracture network inversion method and system

Abstract

The application relates to the technical field of fracture network exploration, and discloses a water-rich fracture network inversion method and system, a detection parameter file is established in advance, a detection method corresponding to each detection batch, a working face advancing record, a spatial position of a detection point, a detection start and end time, a detection depth range and abnormal response data are collected; a multi-source original detection table is generated and advancing phase backtracking processing is carried out to obtain phase correction coordinates unified in a target working face position; an equivalent detection volume unit is constructed based on a sensitive direction, a resolution radius, a detection depth range and the phase correction coordinates, a multi-source abnormal volume table is generated, cross-source attribution scores are calculated, candidate connection edges are established and a water-rich fracture candidate network is obtained. The application can unify abnormal responses corresponding to different detection methods and different detection batches to the same target working face position for comparison, reduce abnormal body mismatch and network misconnection, and provide a basis for connectivity determination and risk classification.

Description

Technical Field

[0001] This application relates to the field of fracture network exploration technology, specifically to a method and system for inverting water-rich fracture networks. Background Technology

[0002] For water-rich anomalies in underground media such as water-rich fissures and water-rich fractured zones, engineering sites typically use different detection methods to acquire data in batches, and then interpret the results and analyze fissure connectivity. Since the spatial location of the measuring points, the time of acquisition, the range of detection depth, and the construction stage differ for each batch of detection, and different detection methods also vary in their sensitivity to the object being detected, spatial resolution, and coverage, current technologies often directly compare, stitch together, and interpret data based on the original measuring point coordinates, single detection mileage, or a single anomaly record.

[0003] However, as the tunnel face advances continuously during construction, the mileage location of the same underground target body changes across different detection batches. If the original location relationships from the initial data collection are still used for judgment, anomalies originally corresponding to the same water-rich fracture segment are easily dispersed into multiple independent anomaly results. Furthermore, the anomaly responses generated by different detection methods are inconsistent in directionality, depth coverage, and lateral resolution. Connecting them solely based on proximity or local overlap can easily misclassify anomalies from different sources and not from the same water-conducting structure as belonging to the same fracture network. Especially in scenarios involving cross-construction cycles, cross-batch, and multi-source data joint interpretation, existing technologies lack effective constraints on the attribution and sequential relationships of underground targets, leading to mismatches, mis-splicing, and distorted connectivity judgments in the inversion results of water-rich fracture networks, thus affecting the identification and risk assessment of water-rich geological bodies.

[0004] Therefore, how to more accurately determine the attribution and network inversion of anomalies corresponding to different detection results under the conditions of multi-source detection, batch acquisition and continuous advancement at the working face has become a technical problem that needs to be solved. Summary of the Invention

[0005] In existing technologies, during geological exploration, geophysical investigation, water-rich anomaly identification, and advanced geological prediction at the tunnel face, anomaly response data from different detection methods, batches, and construction stages are easily affected by the continuous advancement of the tunnel face. This leads to inconsistencies in mileage benchmarks, inconsistent spatial coverage, and unclear cross-source attribution relationships, resulting in data misalignment, incorrect anomaly attribution, and mis-splicing of networks during the inversion of water-rich fracture networks from multi-source detection results. Consequently, these issues hinder the provision of stable data for fracture connectivity determination and risk classification. The purpose of this application is to provide a method and system for inverting water-rich fracture networks. By performing forward phase backtracking processing, equivalent detection volume unit construction, cross-source attribution scoring calculation, and candidate network organization processing on multi-source detection records, this method achieves the effect of unifying anomaly response data from different batches to the same target tunnel face location for comprehensive geophysical interpretation and water-rich fracture identification.

[0006] To achieve the above objectives, this application adopts the following technical solution:

[0007] Firstly, this application provides a method for inverting water-rich fracture networks, including:

[0008] Establish a pre-established detection parameter file; collect detection methods, face advance records, spatial locations of measuring points, detection start and end times, detection depth range, and abnormal response data for each detection batch;

[0009] Based on the sensitive direction and resolution radius in the detection method and detection parameter file corresponding to each detection batch, a multi-source original detection table is generated; based on the face advance record and the start and end times of detection, the multi-source original detection table is subjected to advance phase backtracking processing to obtain phase correction coordinates unified with the target face position;

[0010] Based on the sensitive direction, resolution radius, detection depth range and phase correction coordinates of each detection method in the original multi-source detection table, an equivalent detection volume unit is constructed; the abnormal response data corresponding to the phase correction coordinates is mapped to the equivalent detection volume unit to generate a multi-source anomaly volume table.

[0011] Cross-source attribution scores were calculated based on the multi-source anomaly volume table;

[0012] Candidate connection edges are established based on cross-source attribution scores; the candidate connection edges are then eliminated or retained to obtain a candidate network of water-rich fractures.

[0013] Preferably, the method for generating a multi-source raw detection table includes:

[0014] Based on the detection method corresponding to each detection batch, the corresponding sensitive direction and resolution radius are retrieved from the detection parameter file;

[0015] Based on the tunnel face advancement records and the start and end times of the exploration, the tunnel face mileage and construction cycle number corresponding to each exploration batch are extracted;

[0016] The detection method, working face mileage, construction cycle number, spatial location of measuring point, detection start and end time, detection depth range, abnormal response data, sensitive direction and resolution radius are written into the corresponding information to generate a multi-source original detection table.

[0017] Preferably, the method for obtaining phase correction coordinates uniform to the target face position includes:

[0018] Read the target face position corresponding to the current inversion;

[0019] The advance positions of each probe batch at the start and end times of the probe were determined based on the advance records at the working face.

[0020] Based on the mileage difference between the target face position and the advancing position, an advancing correction amount is generated;

[0021] The spatial position of the measuring point is converted back along the advancing direction of the tunnel face according to the advancing correction amount to obtain the phase correction coordinates unified with the target tunnel face position.

[0022] Preferably, the method for constructing an equivalent probe volume unit includes:

[0023] Based on the sensitive directions corresponding to each detection method in the multi-source original detection table, the extension direction of the equivalent detection volume unit is determined;

[0024] The lateral coverage range of the equivalent detection volume unit is determined based on the resolution radius;

[0025] The longitudinal coverage range of the equivalent detection volume unit is determined based on the detection depth range;

[0026] Using phase-corrected coordinates as the positioning reference for the equivalent detection volume unit, and combining the extension direction, lateral coverage range, and longitudinal coverage range corresponding to the same record in the multi-source original detection table, the equivalent detection volume unit is constructed.

[0027] Preferably, the method for generating a multi-source anomaly volume table includes:

[0028] Based on the phase correction coordinates and the anomaly response spatial interval corresponding to the anomaly response data, the equivalent detection volume unit corresponding to the anomaly response data is determined;

[0029] When the phase correction coordinates fall within the overlapping range of multiple equivalent detection volume units, the target equivalent detection volume unit is determined according to the sensitive direction and detection depth range corresponding to the detection method.

[0030] The equivalent detection volume unit of the target is written to correspond with the anomaly response data to generate a multi-source anomaly volume table.

[0031] Preferably, the method for obtaining cross-source attribution scores includes:

[0032] For the equivalent detection volume units in the multi-source anomaly volume table that originate from different detection methods, the mileage proximity, lateral overlap, azimuth consistency, depth consistency, and construction stage consistency are calculated respectively.

[0033] A cross-source attribution score is formed based on mileage proximity, lateral overlap, orientation consistency, depth consistency, and construction stage consistency.

[0034] Preferably, the methods for calculating mileage proximity, lateral overlap, azimuth consistency, depth consistency, and construction stage consistency include:

[0035] Based on the phase correction coordinates, the mileage difference and lateral overlap relationship between equivalent detection volume units from different detection methods are extracted to form mileage proximity and lateral overlap.

[0036] Based on the correspondence between sensitive directions and detection depth ranges, azimuth consistency and depth consistency are established;

[0037] Based on the correspondence between the construction cycle number and the start and end times of the detection, a consistency of construction stages is established;

[0038] Mileage proximity, lateral overlap, orientation consistency, depth consistency, and construction stage consistency all include first, second, and third levels.

[0039] Preferably, the method for establishing candidate connection edges includes:

[0040] For each equivalent detection volume unit in the multi-source anomaly volume table, cross-source pairing is performed according to the cross-source attribution score from high to low.

[0041] Pairing relationships originating from different detection methods, whose cross-source attribution scores are located in the first or second sorting position of all cross-source pairing results corresponding to the current equivalent detection volume unit, whose mileage proximity and lateral overlap do not belong to the third level, and whose azimuth consistency, depth consistency, and construction stage consistency belong to the third level, are identified as candidate connection edges; otherwise, the pairing relationships are eliminated.

[0042] Preferably, the method for obtaining a water-rich fracture candidate network includes:

[0043] For multiple candidate connection edges corresponding to the same equivalent detection volume unit, sort them according to the cross-source attribution score;

[0044] Candidate connecting edges that are ranked first and originate from different detection methods are determined as priority connecting edges;

[0045] For candidate connection edges other than the priority connection edges, retain or remove them based on the sensitive direction correspondence, cross-source attribution score, construction stage consistency and depth consistency of the candidate connection edges;

[0046] A candidate network of water-rich fractures is formed based on the retained candidate connection edges.

[0047] Preferably, the detection method, phase correction coordinates, and cross-source attribution score corresponding to each candidate connection edge in the water-rich fracture candidate network are associated and recorded for use in connectivity determination and risk classification.

[0048] Secondly, this application provides a water-rich fracture network inversion system, comprising:

[0049] The parameter acquisition module is used to pre-establish a detection parameter file; and to collect the detection method, face advance record, spatial location of measuring point, detection start and end time, detection depth range and abnormal response data corresponding to each detection batch.

[0050] The phase correction module generates a multi-source original detection table based on the sensitive direction and resolution radius in the detection method and detection parameter file corresponding to each detection batch; based on the face advance record and the start and end times of detection, the multi-source original detection table is processed by advance phase backtracking to obtain phase correction coordinates unified with the target face position;

[0051] The volume equivalent module is used to construct an equivalent detection volume unit based on the sensitive direction, resolution radius, detection depth range and phase correction coordinates of each detection method in the multi-source original detection table.

[0052] The attribution scoring module is used to map the anomaly response data corresponding to the phase correction coordinates to the equivalent detection volume unit, and generate a multi-source anomaly volume table; the cross-source attribution score is calculated based on the multi-source anomaly volume table.

[0053] The network inversion module is used to establish candidate connection edges based on cross-source attribution scores; the candidate connection edges are then processed to eliminate or retain them to obtain a candidate network of water-rich fractures.

[0054] Compared with the prior art, the beneficial effects achieved by this application are as follows:

[0055] This application first establishes a pre-built detection parameter file and collects the detection methods, face advance records, spatial locations of measuring points, detection start and end times, detection depth ranges, and anomaly response data corresponding to each detection batch. This enables the original scattered and inconsistent multi-source detection records to be organized into a unified input that can be further processed, thereby reducing the misalignment and misbinding of front-end geological detection data before entering the subsequent processing stage.

[0056] This application further advances the phase backtracking processing of multi-source original detection tables to obtain phase correction coordinates unified at the target face location. By combining the sensitive direction, resolution radius and detection depth range, an equivalent detection volume unit is constructed. This enables the unification of anomalous response data corresponding to different detection methods, different detection batches and different construction stages to the same target face location and the same spatial coverage, thereby improving the comparability of geophysical exploration results in the water-rich fracture inversion process.

[0057] This application further generates a multi-source anomaly volume table by mapping the anomaly response data corresponding to the phase correction coordinates to the equivalent detection volume unit, and calculates the cross-source attribution score based on the multi-source anomaly volume table. This can organize the originally scattered mileage relationship, lateral relationship, orientation relationship, depth relationship and construction stage relationship into an attribution judgment basis that can be directly called, thereby enhancing the reliability of water-rich anomaly identification and anomaly attribution analysis.

[0058] This application further establishes candidate connection edges based on cross-source attribution scores and performs elimination or retention processing on the candidate connection edges to obtain a water-rich fracture candidate network. This can reduce the abnormal attribution errors and network mis-splitting problems caused by connecting based solely on a single coordinate proximity relationship, making the fracture network organization results more suitable for fracture connectivity determination, advanced geological prediction, and risk classification.

[0059] This application also improves the ability to continuously utilize network results and trace results in subsequent calling stages by associating and recording the detection methods, phase correction coordinates and cross-source attribution scores corresponding to each candidate connection edge in the water-rich fracture candidate network.

[0060] In summary, this application does not merely perform routine analysis on single detection results, but rather constructs a closed processing chain around the unified alignment, unified volume representation, unified attribution scoring, and unified network organization of multi-source geophysical detection results under tunnel face advancement conditions. This improves the accuracy, consistency, and stability of the water-rich fracture network inversion results and reduces the risk of misjudgment and mis-splicing of multi-source detection results in complex construction scenarios. Attached Figure Description

[0061] Figure 1 This is a flowchart illustrating the water-rich fracture network inversion method proposed in this application.

[0062] Figure 2 This is a logical diagram illustrating the process of cross-source attribution scoring and candidate connection edge establishment in this application.

[0063] Figure 3 This is a schematic diagram of the structure of a water-rich fracture network inversion system according to this application. Detailed Implementation

[0064] The technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments and specific features in the embodiments are detailed descriptions of the technical solution of this application, rather than limitations thereof.

[0065] Example 1:

[0066] like Figure 1 and Figure 2 As shown, this embodiment provides a method for inverting water-rich fracture networks, specifically including:

[0067] Establish a pre-established detection parameter file; collect the detection method, face advance record, spatial location of measuring point, detection start and end time, detection depth range and abnormal response data corresponding to each detection batch.

[0068] The tunnel face is the working face at the front of the tunnel construction process where excavation, support, or advanced exploration is being carried out. The location of the tunnel face is indicated by the mileage of the tunnel face in the direction of tunnel mileage. The tunnel face advancement record is a data record that records the advancement time, advancement mileage, and construction cycle number in chronological order.

[0069] The method for establishing a detection parameter archive includes: classifying various detection methods involved in subsequent processing and creating an independent archive entry for each detection method; for each detection method, determining the spatial extension direction that preferentially corresponds to the abnormal response based on the extension direction of the survey line, the projection direction, or the drilling direction in the corresponding detection operation, and identifying the spatial extension direction as the sensitive direction; further, for each detection method, statistically analyzing the minimum distance between adjacent measurement points in the historical detection batch results corresponding to the same detection method when the abnormal responses can be identified separately, where the historical detection batch results are the records corresponding to detection batches that have been completed and verified by excavation; and determining the resolution radius. First, in the historical detection results corresponding to the same detection method, the abnormal response data corresponding to two adjacent measuring points are read in groups. When the abnormal response data corresponding to two adjacent measuring points are within the same detection depth range and the start and end positions do not coincide, or when the abnormal response data corresponding to two adjacent measuring points are within different detection depth ranges, the abnormal response data corresponding to the two adjacent measuring points are determined to be identifiable separately. Then, the measuring point distance between all identifiable adjacent measuring points is counted, and half of the smallest measuring point distance is determined as the resolution radius. When the same detection method corresponds to multiple historical detection results, the same resolution radius is uniformly used and written into the detection parameter file.

[0070] The method for collecting data on detection methods, face advancement records, spatial locations of measuring points, detection start and end times, detection depth ranges, and abnormal response data for each detection batch includes: using detection batches as the collection unit, data is immediately collected after each detection batch is completed; the detection method is read from the detection operation record; the face advancement record corresponding to the detection start and end times is read from the construction advance record; the spatial location of the measuring point is the batch relative coordinate established with the advancement position of the corresponding detection batch as the origin; the detection start and end times are read from the detection equipment acquisition log; the detection depth range is read from the detection task record and detection result record; abnormal response data corresponding to the spatial location of the measuring point and the detection depth range is extracted from the detection result record; when a detection batch contains multiple spatial locations of measuring points or multiple detection depth ranges, they are written one by one according to the actual acquisition order to ensure that the detection method, face advancement record, spatial location of measuring points, detection start and end times, detection depth range, and abnormal response data in the same acquisition record belong to the same acquisition of the same detection batch; after the acquisition is completed, missing fields, fields with incorrect time order, and fields with incorrect correspondence are reviewed, and only acquisition results with complete fields and consistent correspondence are retained.

[0071] The abnormal response data includes at least one of the following: abnormal response location, abnormal response intensity, abnormal response start depth, abnormal response end depth, and abnormal response category. The abnormal response location is read from the abnormal location field corresponding to the spatial location of the measuring point in the detection result record. The abnormal response intensity is read from the abnormal amplitude, abnormal energy, or abnormal probability fields in the detection result record. The abnormal response start depth and abnormal response end depth are read from the start and end depths of the corresponding abnormal intervals in the detection result record. The abnormal response category is read from the water-rich anomaly, fracture zone anomaly, or fissure anomaly markings in the detection result record. When the data types output by different detection methods are inconsistent, the data types output by different detection methods are uniformly mapped to the phase correction coordinates and detection depth range for recording.

[0072] The reason for collecting the above data is that the detection method is used to call the corresponding sensitive direction and resolution radius from the detection parameter file to ensure that the abnormal response data corresponding to different detection methods have a consistent calling basis in subsequent processing; the tunnel face advance record and detection start and end time are used to determine the front-to-back relationship of each detection batch relative to the target tunnel face position and to provide a time and mileage basis for advance phase backtracking processing; the spatial location of the measuring point is used to give the initial position of the abnormal response data in the tunnel space, the detection depth range is used to limit the forward coverage interval corresponding to the abnormal response data, and the abnormal response data is used to characterize the abnormal distribution within the corresponding coverage interval; only by simultaneously collecting the detection method, tunnel face advance record, measuring point spatial location, detection start and end time, detection depth range, and abnormal response data can a multi-source original detection table with complete fields and clear correspondence be generated, and on the basis of the multi-source original detection table, the advance phase backtracking processing, equivalent detection volume unit construction, multi-source abnormal volume table generation, and cross-source attribution score calculation can be completed.

[0073] Based on the sensitive direction and resolution radius in the detection method and detection parameter file corresponding to each detection batch, a multi-source original detection table is generated.

[0074] Methods for generating multi-source raw probe tables include:

[0075] Based on the detection method corresponding to each detection batch, the corresponding sensitive direction and resolution radius are retrieved from the detection parameter file; based on the face advance record and detection start and end times, the face mileage and construction cycle number corresponding to each detection batch are extracted; the detection method, face mileage, construction cycle number, spatial location of measuring point, detection start and end times, detection depth range, abnormal response data, sensitive direction and resolution radius are written into the corresponding data to generate a multi-source original detection table.

[0076] Specifically, the face mileage and construction cycle number corresponding to each detection batch are as follows: The advance position corresponding to the start and end times of the detection is read from the face advance record; when the start and end times of the detection are within the same construction cycle, the next advance position among the advance positions corresponding to the start and end times of the detection is determined as the face mileage, and the cycle number in the construction record corresponding to the start and end times of the detection is determined as the construction cycle number; when the start and end times of the detection span two adjacent construction cycles, the starting advance position of the next construction cycle among the advance positions corresponding to the start and end times of the detection is determined as the face mileage, and the cycle number corresponding to the next construction cycle is determined as... Construction cycle number; when there are multiple spatial locations of measuring points or multiple detection depth ranges within a detection batch, all collected records within the same detection batch are uniformly mapped to the same working face mileage and the same construction cycle number to ensure that the time assignment and construction stage assignment of the same detection batch in the multi-source original detection table are consistent; after completing the extraction of the working face mileage and construction cycle number, the working face mileage and construction cycle number are then written into the multi-source original detection table one by one with the detection method, spatial location of measuring points, detection start and end time, detection depth range, abnormal response data, sensitive direction and resolution radius corresponding to the same detection batch.

[0077] By generating a multi-source original detection table, the detection methods, face mileage, construction cycle number, spatial location of measuring points, detection start and end times, detection depth range, abnormal response data, sensitive direction, and resolution radius, which were originally scattered across different detection batches, can be uniformly organized into a single table. On the one hand, the multi-source original detection table provides a basis for subsequent advance phase backtracking processing to perform corresponding conversions based on the detection start and end times and face advance records, avoiding confusion in the correspondence between different detection batches at the advance positions. On the other hand, the multi-source original detection table establishes a fixed correspondence between abnormal response data and sensitive direction, resolution radius, spatial location of measuring points, and detection depth range, enabling the direct use of relevant fields from the same record when constructing equivalent detection volume units, reducing field mismatches during cross-table extraction. Furthermore, by introducing the construction cycle number into the multi-source original detection table, the abnormal response data corresponding to different detection batches can be limited to a specific construction stage, providing a unified data foundation for subsequent calculations of construction stage consistency and cross-source attribution scores.

[0078] Based on the advance records and the start and end times of the probes at the tunnel face, the advance phase backtracking process is performed on the original multi-source probe tables to obtain phase-corrected coordinates that are consistent with the target tunnel face position.

[0079] To ensure that the spatial location, abnormal response location, and detection depth range of the measuring points corresponding to different detection batches can be converted under the same coordinate reference, this embodiment first establishes a coordinate system in front of the tunnel face before performing the advance phase backtracking process. The coordinate system in front of the tunnel face takes the projection point of the target tunnel face position in the fixed tunnel coordinates as the origin, the advance direction of the tunnel face as the longitudinal axis, the positive direction of the longitudinal axis pointing from the target tunnel face position to the area to be detected in front of the tunnel, the horizontal lateral direction perpendicular to the advance direction of the tunnel face as the transverse axis, the positive direction of the transverse axis determined according to the right side direction of the tunnel in the construction survey record, and the elevation direction as the elevation axis, the positive direction of the elevation axis determined according to the vertical upward direction in the construction survey record. The fixed tunnel coordinates are the coordinates expressed in the construction survey record as the tunnel design mileage, horizontal lateral position, and elevation position. The spatial location of the measuring points is stored in the multi-source original detection table in the form of batch relative coordinates. The batch relative coordinates take the advance position of the corresponding detection batch as the longitudinal zero point and are expressed as the longitudinal distance, horizontal lateral distance, and elevation distance of the measuring point relative to the advance position of the corresponding detection batch.

[0080] When the spatial location of the measuring point in the measuring point layout record is a fixed tunnel coordinate, first read the advance position of the corresponding detection batch, subtract the mileage value corresponding to the advance position from the mileage value in the measuring point spatial location to obtain the batch relative longitudinal value, and retain the horizontal lateral value and elevation value in the measuring point spatial location to form the batch relative coordinate; when the spatial location of the measuring point in the measuring point layout record is already the batch relative coordinate, directly write the batch relative coordinate into the multi-source original detection table.

[0081] The method for converting the abnormal response location and detection depth range into an abnormal response spatial interval at the target face location is as follows: First, the spatial position of the measuring point is converted into phase-corrected coordinates based on the advance correction amount. Then, the phase-corrected coordinates are used as the starting position of the abnormal response spatial interval. The abnormal response location, the starting depth of the abnormal response, the ending depth of the abnormal response, and the detection depth range are read along the sensitive direction corresponding to the detection method. When the abnormal response data contains both the starting depth and the ending depth of the abnormal response, the starting depth and the ending depth of the abnormal response are used as the starting and ending boundaries of the abnormal response spatial interval in the sensitive direction. When the abnormal response data does not contain the starting depth and the ending depth of the abnormal response separately, the starting depth and the ending depth of the detection depth range are used as the starting and ending boundaries of the abnormal response spatial interval in the sensitive direction.

[0082] The start and end boundaries of the anomaly response space interval are then projected onto the longitudinal axis, transverse axis, and elevation axis of the coordinate system in front of the tunnel face to obtain an anomaly response space interval unified with the target tunnel face position. The subsequent mileage difference is determined based on the coordinate difference between the two phase correction coordinates on the longitudinal axis, and the lateral overlap is determined based on the coincidence relationship of the transverse coverage range of the two equivalent detection volume units on the transverse axis.

[0083] Methods for obtaining phase correction coordinates uniform to the target face position include:

[0084] Read the target face position corresponding to the current inversion; determine the advance position of each detection batch at the start and end times of detection based on the face advance record; form the advance correction amount based on the mileage difference between the target face position and the advance position; back-calculate the spatial position of the measuring point along the advance direction of the face according to the advance correction amount to obtain the phase correction coordinates unified with the target face position.

[0085] The method for determining the advance position of each detection batch at the start and end times of detection based on the face advance record includes: first, reading the start and end times of detection for each detection batch in the multi-source original detection table; then, reading the advance times and corresponding advance mileages continuously recorded in the face advance record in chronological order; when the start and end times of detection are consistent with the advance times in the face advance record, the advance mileage corresponding to the advance time is directly determined as the advance position; when the start and end times of detection are between two adjacent advance times, the advance mileage corresponding to the later advance time is determined as the advance position; when a detection batch corresponds to multiple spatial locations of measuring points or multiple detection depth ranges, the advance position corresponding to the same detection batch is uniformly determined as the advance mileage corresponding to the later time in the detection start and end times; after determination, the advance position is kept in one-to-one correspondence with the detection method, spatial location of measuring points, start and end times of detection, detection depth range, abnormal response data, sensitive direction, and resolution radius of the corresponding detection batch in the multi-source original detection table.

[0086] The method for generating the propulsion correction amount based on the mileage difference between the target face position and the advance position includes: reading the target face position corresponding to the current inversion; then reading the advance position corresponding to each detection batch; subtracting the advance position mileage value corresponding to the detection batch from the mileage value corresponding to the target face position to obtain the propulsion correction amount corresponding to the detection batch. The calculation method is as follows:

[0087] ;

[0088] in, This represents the propulsion correction amount corresponding to the r-th multi-source original detection table record; This indicates the mileage value corresponding to the target working face position. Read from the target face position corresponding to the current inversion; This indicates that the r-th multi-source raw detection table record contains the propulsion position mileage value corresponding to the detection batch. Read from the face advance record based on the start and end times of the probe; when When, it indicates that the target face is located ahead of the advance position corresponding to the detection batch; when When, it indicates that the target face position is consistent with the advancement position corresponding to the detection batch; when When the target face is located behind the corresponding advance position of the detection batch, the advance correction amount is generated and written into the multi-source original detection table one by one, so that each record in the multi-source original detection table is processed by advance phase backtracking according to the corresponding advance correction amount.

[0089] The method for determining the target working face position during the current inversion is as follows: First, read the current inversion start time; then, read the continuously recorded advancement times and corresponding advancement mileages in the working face advancement record in chronological order; when the current inversion start time is consistent with the advancement time in the working face advancement record, directly determine the corresponding advancement mileage as the target working face position; when the current inversion start time is between two adjacent advancement times, determine the advancement mileage corresponding to the later advancement time as the target working face position; when the current inversion start time spans a construction cycle switch, determine the first advancement mileage corresponding to the next construction cycle as the target working face position.

[0090] The method for converting the spatial location of the measuring point into phase-corrected coordinates uniform with the target face position according to the advance correction amount includes: reading the spatial location and advance correction amount of the measuring point corresponding to the r-th record in the multi-source original detection table; the spatial location of the measuring point is represented by batch relative coordinates, which include batch relative longitudinal value, batch relative lateral value, and batch relative elevation value; the batch relative longitudinal value takes the advance position corresponding to the detection batch to which the r-th record belongs as the longitudinal zero point, and the batch relative lateral value and batch relative elevation value are taken along the lateral axis and elevation axis of the coordinate system in front of the face, respectively; the phase-corrected coordinates include phase-corrected longitudinal coordinate values, phase-corrected lateral coordinate values, and phase-corrected elevation coordinate values; for example, the phase-corrected longitudinal coordinate value can be determined according to the following formula:

[0091] ;

[0092] For example, the phase-corrected lateral coordinate value and the phase-corrected elevation coordinate value can be determined according to the following formula:

[0093] ;

[0094] ;

[0095] in, This represents the phase-corrected longitudinal coordinate value corresponding to the r-th multi-source original detection table record; This represents the phase-corrected lateral coordinate value corresponding to the r-th multi-source original detection table record; This represents the phase-corrected elevation coordinate value corresponding to the r-th multi-source original detection table record; This represents the batch relative longitudinal value corresponding to the r-th multi-source raw detection table record. Read the batch relative coordinates from the spatial location of the measurement points; This represents the batch-wise relative horizontal value corresponding to the r-th multi-source original detection table record. Read the batch relative coordinates from the spatial location of the measurement points; This represents the batch relative elevation value corresponding to the r-th multi-source original detection table record. Read the batch relative coordinates from the spatial location of the measurement points; This represents the propulsion correction amount corresponding to the r-th multi-source original detection table record; when At that time, the target face was located ahead of the advance position corresponding to the detection batch. Less than ;when hour, equal ;when At that time, the target face was located behind the advance position corresponding to the detection batch. Greater than After the conversion is completed, , and The phase correction coordinates are combined and written into the multi-source original detection table, so that the phase correction coordinates correspond to the detection method, detection depth range, abnormal response data, sensitive direction and resolution radius in the same record.

[0096] When it is necessary to convert the location of the abnormal response and the detection depth range into the spatial interval of the abnormal response at the target face location, first read the r-th original multi-source detection table record of the corresponding phase correction coordinates, sensitive direction, abnormal response start depth, and abnormal response end depth; when the abnormal response data does not separately record the abnormal response start depth and abnormal response end depth, the start depth and end depth in the detection depth range are used as the abnormal response start depth and abnormal response end depth, respectively; convert the sensitive direction into a unit direction vector in the coordinate system in front of the face, and the unit direction vector can be represented in the following form:

[0097] ;

[0098] in, This represents the sensitive direction unit vector corresponding to the r-th multi-source original detection table record; This represents the component of the sensitive direction unit vector on the vertical axis; This represents the component of the sensitive direction unit vector on the horizontal axis; This represents the component of the sensitive direction unit vector on the elevation axis; The sensitive directions are generated based on the sensitive directions recorded in the detection parameter archive. The generation method is to decompose the sensitive directions in the coordinate system in front of the tunnel face and then normalize them.

[0099] The starting and ending coordinates of the anomaly response space interval can be determined according to the following formula:

[0100] ;

[0101] ;

[0102] in, This represents the coordinates of the starting point of the anomaly response space interval corresponding to the r-th multi-source original detection table record; This represents the coordinates of the endpoint of the anomaly response space interval corresponding to the r-th multi-source original detection table record; This indicates the starting depth of the abnormal response corresponding to the r-th multi-source original detection table record; This indicates the termination depth of the abnormal response corresponding to the r-th multi-source original detection table record; and Read from the anomaly response data; if the anomaly response start depth and anomaly response end depth are not recorded in the anomaly response data, and Read from the detection depth range; the coordinates of the start and end points of the anomaly response spatial interval together define an anomaly response spatial interval unified at the target face position; the subsequent longitudinal coverage range is based on the anomaly response spatial interval in the sensitive direction. to Confirmed; the mileage difference is based on the two phase correction coordinates. The difference is determined; the lateral overlap is based on the two equivalent probe volume units. The overlap of lateral coverage in the direction is determined.

[0103] By obtaining phase-corrected coordinates unified to the target face position, the spatial positions of measurement points collected by different detection batches at different advance times can be uniformly converted to the same target face position for expression. On the one hand, this reduces the accumulation of misalignment of the same abnormal response data in the mileage direction caused by continuous face advance, enabling the abnormal response data in the subsequent multi-source abnormal volume table to be mapped under a unified advance benchmark. On the other hand, after the phase-corrected coordinates maintain a fixed correspondence with the detection method, detection depth range, abnormal response data, sensitive direction, and resolution radius, the spatial landing point corresponding to the abnormal response data can be directly determined based on the phase-corrected coordinates when constructing the equivalent detection volume unit. This provides a unified coordinate basis for cross-source attribution scoring, candidate connection edge establishment, and water-rich fracture candidate network generation.

[0104] Based on the sensitive direction, resolution radius, detection depth range and phase correction coordinates corresponding to each detection method in the multi-source original detection table, an equivalent detection volume unit is constructed.

[0105] Methods for constructing equivalent probe volume units include:

[0106] Based on the sensitive directions corresponding to each detection method in the multi-source original detection table, the extension direction of the equivalent detection volume unit is determined; the lateral coverage of the equivalent detection volume unit is determined based on the resolution radius; the longitudinal coverage of the equivalent detection volume unit is determined based on the detection depth range; the phase correction coordinates are used as the positioning reference of the equivalent detection volume unit, and the equivalent detection volume unit is constructed by combining the extension direction, lateral coverage, and longitudinal coverage of the same record in the multi-source original detection table.

[0107] The method for determining the extension direction of the equivalent detection volume unit based on the sensitive directions corresponding to each detection method in the multi-source original detection table includes: first, reading the detection method and its corresponding sensitive direction in the multi-source original detection table line by line; then checking whether the sensitive direction is consistent with the detection deployment direction of the corresponding detection method; when the sensitive direction is consistent with the detection deployment direction of the corresponding detection method, directly determining the sensitive direction as the extension direction of the equivalent detection volume unit; when the sensitive direction is inconsistent with the detection deployment direction of the corresponding detection method, correcting the sensitive direction with the actual detection deployment direction corresponding to the detection method, and determining the corrected sensitive direction as the extension direction of the equivalent detection volume unit; after determination, maintaining a line-by-line correspondence between the extension direction and the phase correction coordinates, resolution radius, and detection depth range in the corresponding records to ensure that the equivalent detection volume units corresponding to different detection methods all extend along the direction of the actual perceived abnormal response of the corresponding detection method.

[0108] The method for determining the lateral coverage of an equivalent detection volume unit based on the resolution radius includes: first, reading the resolution radius corresponding to each record in the multi-source original detection table; then, taking the extension direction as the center direction, determining the lateral boundaries on both sides of the extension direction according to the resolution radius; when different detection methods correspond to different resolution radii, independently determining the lateral coverage according to the resolution radius in the corresponding record; when there are multiple phase correction coordinates in the same detection batch, calling the resolution radius in the corresponding record for each phase correction coordinate to determine the lateral coverage; after determination, maintaining the lateral coverage with the extension direction, phase correction coordinates, and detection depth range in the corresponding record, so that the lateral coverage can characterize the lateral coverage range of the resolvable abnormal response of the corresponding detection method near the current phase correction coordinate.

[0109] The method for determining the longitudinal coverage of an equivalent detection volume unit based on the detection depth range includes: reading the detection depth range corresponding to each record in the multi-source original detection table; extracting the starting depth and ending depth from the detection depth range; determining the starting depth as the starting boundary of the longitudinal coverage range and the ending depth as the ending boundary of the longitudinal coverage range; when the detection depth range corresponds to multiple consecutive depth intervals, merging the multiple consecutive depth intervals into the same longitudinal coverage range in the order of the records; when the detection depth range corresponds to only a single depth interval, directly determining the longitudinal coverage range using the single depth interval; after determination, maintaining a line-by-line correspondence between the longitudinal coverage range and the phase correction coordinates, extension direction, and lateral coverage range, so that the longitudinal coverage range can characterize the depth interval covered forward along the extension direction by the anomalous response.

[0110] The method for constructing an equivalent detection volume unit by combining phase correction coordinates, extension direction, lateral coverage range, and longitudinal coverage range includes: first, reading the phase correction coordinates, extension direction, lateral coverage range, and longitudinal coverage range corresponding to the same multi-source original detection table record; then, determining the phase correction coordinates as the positioning reference for the equivalent detection volume unit; starting from the phase correction coordinates, extending forward along the extension direction according to the longitudinal coverage range, and then expanding to both sides according to the lateral coverage range within the entire expanded range, thereby forming an equivalent detection volume unit corresponding to the current record; when there are multiple records in the multi-source original detection table, repeating the above construction process one by one according to the record order, and ensuring that each equivalent detection volume unit corresponds to only one multi-source original detection table record in terms of detection method, phase correction coordinates, detection depth range, abnormal response data, sensitive direction, and resolution radius; after the construction is completed, each equivalent detection volume unit is written into the subsequent processing results for mapping of the abnormal response data corresponding to the phase correction coordinates.

[0111] By constructing equivalent detection volume units, the anomaly response data corresponding to different detection methods, which are recorded using scattered phase correction coordinates and detection depth ranges, can be transformed into a volumetric representation with unified spatial coverage. On the one hand, the equivalent detection volume unit incorporates the sensitive direction, resolution radius, and detection depth range into the same construction process, enabling the anomaly response data corresponding to different detection methods to have a consistent spatial carrying capacity when compared subsequently, reducing mismatches caused by direct comparison based solely on phase correction coordinates. On the other hand, after the equivalent detection volume unit maintains a fixed correspondence with the detection method, phase correction coordinates, anomaly response data, sensitive direction, and resolution radius in the corresponding records, the spatial coverage range corresponding to the anomaly response data can be directly determined when mapping the anomaly response data corresponding to the phase correction coordinates to the equivalent detection volume unit. This provides a unified volumetric basis for generating multi-source anomaly volume tables, calculating cross-source attribution scores, and establishing candidate connection edges.

[0112] The abnormal response data corresponding to the phase correction coordinates are mapped to the equivalent detection volume unit to generate a multi-source abnormal volume table.

[0113] Methods for generating multi-source anomaly volume tables include:

[0114] Based on the phase correction coordinates and the anomaly response space interval corresponding to the anomaly response data, the equivalent detection volume unit corresponding to the anomaly response data is determined; when the phase correction coordinates fall within the overlapping range of multiple equivalent detection volume units, the target equivalent detection volume unit is determined according to the sensitive direction and detection depth range corresponding to the detection method; the target equivalent detection volume unit and the anomaly response data are written into the table to generate a multi-source anomaly volume table.

[0115] The multi-source anomaly volume table includes at least the detection batch identifier, detection method, phase correction coordinates, target equivalent detection volume unit identifier, extension direction, lateral coverage, longitudinal coverage, anomaly response data, construction cycle number, and detection start and end times. The detection batch identifier is read from the batch number in the detection operation record. When the batch number is not recorded in the detection operation record, the detection batch identifier is generated by combining the detection method, detection start and end times, and construction cycle number. The target equivalent detection volume unit identifier is generated according to the construction order of the equivalent detection volume units and maintains a one-to-one correspondence with the target equivalent detection volume units.

[0116] Based on the spatial intervals of the abnormal response data corresponding to the phase correction coordinates and the abnormal response data, the equivalent detection volume unit corresponding to the abnormal response data is determined. Specifically, the phase correction coordinates, abnormal response data, and the spatial coverage range corresponding to all equivalent detection volume units are read one by one. Then, the phase correction coordinates are compared with the spatial coverage range corresponding to each equivalent detection volume unit. When the phase correction coordinates are within the extended direction coverage range, lateral coverage range, and longitudinal coverage range corresponding to a certain equivalent detection volume unit, the abnormal response data is initially mapped to that equivalent detection volume unit. When the phase correction coordinates are only within the spatial coverage range of one equivalent detection volume unit, the abnormal response data is directly mapped to that equivalent detection volume unit. The equivalent detection volume unit is then identified as the target equivalent detection volume unit corresponding to the abnormal response data. When the phase correction coordinates do not fall within the spatial coverage of any equivalent detection volume unit, the sensitive direction, resolution radius, and detection depth range in the corresponding record are read first, and the corresponding equivalent detection volume unit is reconstructed using the phase correction coordinates as the positioning reference. When the reconstructed phase correction coordinates fall within the reconstructed equivalent detection volume unit, the abnormal response data is written into the multi-source abnormal volume table. When the reconstructed phase correction coordinates still do not fall within the reconstructed equivalent detection volume unit, the corresponding abnormal response data is not written into the multi-source abnormal volume table, and the corresponding mapping record is deleted.

[0117] The overlap range of multiple equivalent detection volume units is specifically defined as follows: Two or more equivalent detection volume units simultaneously have a shared coverage area at the same spatial location, forming a spatial overlap range; when multiple equivalent detection volume units overlap in their extension direction coverage, lateral coverage, and longitudinal coverage, the portion where all three overlap is defined as the overlap range of the multiple equivalent detection volume units; when only the extension direction coverage overlaps, or only the lateral coverage overlaps, or only the longitudinal coverage overlaps, and the other coverage areas do not overlap, the corresponding portion is not defined as the overlap range of the multiple equivalent detection volume units; only when the phase correction coordinates simultaneously fall within the spatial overlap portion formed by the extension direction coverage, lateral coverage, and longitudinal coverage of multiple equivalent detection volume units is the phase correction coordinate considered to fall within the overlap range of the multiple equivalent detection volume units.

[0118] When the phase correction coordinates fall within the overlapping range of multiple equivalent detection volume units, the method for determining the target equivalent detection volume unit based on the sensitive direction and detection depth range corresponding to the detection method includes: reading the detection method, sensitive direction, and detection depth range from the record corresponding to the phase correction coordinates; then reading the extension direction and longitudinal coverage range corresponding to the multiple equivalent detection volume units falling within the overlapping range; prioritizing the retention of equivalent detection volume units whose extension direction is consistent with the sensitive direction corresponding to the detection method, and removing equivalent detection volume units whose extension direction is inconsistent with the sensitive direction corresponding to the detection method.

[0119] If multiple equivalent detection volume units still exist after filtering based on the sensitive direction, the correspondence between the longitudinal coverage range and the detection depth range of the multiple equivalent detection volume units is compared, and equivalent detection volume units whose starting and ending positions of the longitudinal coverage range both fall within the detection depth range are retained. If multiple equivalent detection volume units still meet both the sensitive direction and detection depth range requirements, the overlap length ratio between the longitudinal coverage range and the detection depth range of each equivalent detection volume unit is calculated. The overlap length ratio is the ratio obtained by dividing the overlap length of the longitudinal coverage range and the detection depth range on the same depth axis by the length of the detection depth range. The equivalent detection volume unit with the largest overlap length ratio is determined as the target equivalent detection volume unit. When at least two equivalent detection volume units have the same overlap length ratio, the directional angle between the extension direction of each equivalent detection volume unit and the sensitive direction is calculated, and the equivalent detection volume unit with the smallest directional angle is determined as the target equivalent detection volume unit.

[0120] After the screening is completed, the final retained equivalent detection volume unit is determined as the target equivalent detection volume unit, and the target equivalent detection volume unit and the corresponding anomaly response data are written into the multi-source anomaly volume table.

[0121] By generating a multi-source anomaly volume table, a fixed correspondence can be established between the anomaly response data corresponding to the phase correction coordinates and the target equivalent detection volume unit. Furthermore, anomaly response data from different detection batches and methods can be uniformly converted into directly comparable volume records. On one hand, the multi-source anomaly volume table ensures that anomaly response data is no longer confined to scattered measurement point locations and detection result records, but is instead mapped to corresponding equivalent detection volume units, facilitating subsequent cross-source attribution analysis based on a unified space. On the other hand, the multi-source anomaly volume table completes the identification of overlapping ranges and the screening of target equivalent detection volume units during the writing process, reducing the possibility of the same anomaly response data being repeatedly mapped to multiple equivalent detection volume units. This provides a consistent and spatially clear data foundation for subsequent calculation of cross-source attribution scores, establishment of candidate connection edges, and generation of water-rich fracture candidate networks.

[0122] Cross-source attribution scores were calculated based on the multi-source anomaly volume table.

[0123] Methods for obtaining cross-source attribution scores include:

[0124] For the equivalent detection volume units in the multi-source anomaly volume table that originate from different detection methods, the mileage proximity, lateral overlap, azimuth consistency, depth consistency, and construction stage consistency are calculated respectively.

[0125] A cross-source attribution score is formed based on mileage proximity, lateral overlap, orientation consistency, depth consistency, and construction stage consistency.

[0126] Among them, mileage proximity, lateral overlap, azimuth consistency, depth consistency, and construction stage consistency are all represented by first, second, and third levels. The first level indicates that the correspondence between two equivalent detection volume units can be given priority for cross-source attribution; the second level indicates that the correspondence between two equivalent detection volume units can participate in candidate screening; and the third level indicates that the correspondence between two equivalent detection volume units is not used as a priority connection basis. The first, second, and third levels are read from the preset level table according to the judgment rules corresponding to each indicator. The preset level table is established based on the pairing records of equivalent detection volume units that have been confirmed by the excavation results to belong to the same water-rich fracture segment in the historical detection batch results, as well as the pairing records of equivalent detection volume units that have been confirmed by the excavation results not to belong to the same water-rich fracture segment.

[0127] Furthermore, the method for generating cross-source attribution scores includes: converting mileage proximity, lateral overlap, azimuth consistency, depth consistency, and construction stage consistency into corresponding level codes; the first level corresponds to level code 3, the second level to level code 2, and the third level to level code 0; for the cross-source pairing relationship formed by the r-th equivalent detection volume unit and the q-th equivalent detection volume unit, the cross-source attribution score can be calculated according to the following formula:

[0128] ;

[0129] in, This represents the cross-source attribution score between the r-th equivalent detector volume cell and the q-th equivalent detector volume cell; This indicates the level code corresponding to the mileage proximity. This indicates the level code corresponding to the lateral overlap. The grade code representing the degree of orientation consistency; This indicates the level code corresponding to the depth consistency. This indicates the grade code corresponding to the consistency of the construction phases; Indicates the weight of the indicator corresponding to mileage proximity; Indicates the index weight corresponding to lateral overlap; The weight of the indicator corresponding to the degree of orientation consistency; This indicates the weight of the metric corresponding to the depth of consistency. This indicates the weight of the indicators corresponding to the consistency of the construction phase.

[0130] For example, the indicator weights satisfy the following relationship:

[0131] ;

[0132] ;

[0133] in, , , , and The settings are based on historical survey batch results. The method involves first reading paired records of equivalent survey volume units that have been confirmed by excavation results to belong to the same water-rich fracture segment, and paired records of equivalent survey volume units that have been confirmed by excavation results not to belong to the same water-rich fracture segment. Then, the distinguishing effects of mileage proximity, lateral overlap, azimuth consistency, depth consistency, and construction stage consistency between correctly paired and incorrectly paired records are calculated. The greater the distinguishing effect, the greater the corresponding indicator weight. The distinguishing effect is determined by the difference between the proportion of the corresponding indicator appearing at the first level in correctly paired records and the proportion appearing at the first level in incorrectly paired records; the larger the difference, the greater the distinguishing effect. When the historical survey batch results are insufficient to distinguish the weights of each indicator, [the following steps are taken]. and Set to not less than , and To ensure that mileage proximity and lateral overlap serve as the basis for spatial correspondence, priority is given to participating in cross-source attribution scoring calculations.

[0134] After calculating the cross-source attribution scores, the cross-source pairings are sorted from highest to lowest score. When two cross-source pairings have the same cross-source attribution score, they are compared first. The value of, and Cross-source pairings with larger values ​​are ranked higher in the sort order; when If the values ​​are still the same, then compare. The value of, and Cross-source pairings with larger values ​​are ranked higher in the sort order; when When the values ​​are still the same, the sorting position is determined according to the record order in the multi-source anomaly volume table; the cross-source attribution score is used for subsequent cross-source pairing sorting, candidate connection edge screening and water-rich fracture candidate network generation.

[0135] Methods for establishing mileage proximity, lateral overlap, orientation consistency, depth consistency, and construction stage consistency include:

[0136] Based on the phase correction coordinates, the mileage difference and lateral overlap relationship between equivalent detection volume units from different detection methods are extracted to form mileage proximity and lateral overlap; based on the correspondence between sensitive direction and detection depth range, azimuth consistency and depth consistency are formed; based on the correspondence between construction cycle number and detection start and end time, construction stage consistency is formed.

[0137] Based on the phase correction coordinates, the mileage difference and lateral overlap relationship between equivalent detection volume units from different detection methods are extracted to form mileage proximity and lateral overlap. Specifically, two equivalent detection volume units from different detection methods are selected from the multi-source anomaly volume table in the order of recording, and the phase correction coordinates and lateral coverage range corresponding to the two equivalent detection volume units are read respectively. Then, the mileage difference is extracted based on the positional difference of the two phase correction coordinates in the tunnel face advance direction. The maximum value between the resolution radius corresponding to the previous equivalent detection volume unit and the resolution radius corresponding to the next equivalent detection volume unit is used as the mileage proximity allowable value, and the mileage proximity allowable range is formed based on the mileage proximity allowable value. When the mileage difference falls within the mileage proximity allowable range, the mileage proximity is written to the first level; when the mileage difference exceeds the mileage... When the distance is close to the allowable range but still within the range that can be continuously covered by the detection depth range corresponding to the two equivalent detection volume units, the mileage proximity is written into the second level; when the mileage difference exceeds the range that can be continuously covered by the detection depth range corresponding to the two equivalent detection volume units, the mileage proximity is written into the third level; then, the lateral overlap relationship is extracted based on the coverage position relationship of the two equivalent detection volume units in the direction perpendicular to the tunnel face advance direction; when the lateral coverage range of the two equivalent detection volume units has a completely overlapping range, the lateral overlap is written into the first level; when the lateral coverage range of the two equivalent detection volume units has only a partially overlapping range, the lateral overlap is written into the second level; when the lateral coverage range of the two equivalent detection volume units has no overlapping range, the lateral overlap is written into the third level.

[0138] Among them, the continuously covered interval is the interval where the union of two detection depth ranges has no break interval on the depth axis; when the interval between two detection depth ranges is not greater than the larger of the two resolution radii, it is considered that there is no break interval.

[0139] After extraction, the mileage proximity and lateral overlap are paired and recorded with the corresponding two equivalent detection volume units.

[0140] Based on the correspondence between sensitive directions and detection depth ranges, azimuth consistency and depth consistency are established. Specifically, two equivalent detection volume units are read to determine the sensitive direction and detection depth range of the detection method, respectively. Then, the azimuth angle between the two sensitive directions is calculated in the coordinate system in front of the tunnel face. The coordinate system in front of the tunnel face has the tunnel face position as the origin, the tunnel face advancement direction as the longitudinal axis, the horizontal lateral direction perpendicular to the tunnel face advancement direction as the transverse axis, and the elevation direction as the vertical axis. When the azimuth angle is not greater than the preset azimuth consistency threshold, the azimuth consistency is written into the first level. When the azimuth angle is greater than the preset azimuth consistency threshold and the front coverage areas corresponding to the two sensitive directions overlap, the azimuth consistency is written into the second level. When the front coverage areas corresponding to the two sensitive directions do not overlap, the azimuth consistency is written into the third level. The preset azimuth consistency threshold is determined based on the upper limit of the azimuth angle between two sensitive directions that have been confirmed by excavation results to belong to the same water-rich fracture segment in historical detection batches.

[0141] The coverage relationship between the start and end positions of the two detection depth ranges is then compared. When there is a continuous overlapping interval between the two detection depth ranges, and the ratio of the length of the overlapping interval to the length of the shorter detection depth range is not less than the preset depth overlap ratio threshold, the depth consistency is written into the first level. When there is a continuous overlapping interval between the two detection depth ranges, and the ratio of the length of the overlapping interval to the length of the shorter detection depth range is less than the preset depth overlap ratio threshold, the depth consistency is written into the second level. When there is no overlapping interval between the two detection depth ranges, the depth consistency is written into the third level. The preset depth overlap ratio threshold is determined based on the lower limit of the overlap ratio of two detection depth ranges that have been confirmed by excavation results to belong to the same water-rich fracture segment in historical detection batches.

[0142] After the comparison is completed, the azimuth and depth consistency are paired and recorded with the corresponding two equivalent detection volume units, and used as the directional and depth basis for the subsequent cross-source attribution score.

[0143] Based on the correspondence between the construction cycle number and the detection start and end times, the construction stage consistency is determined. Specifically, the construction cycle number and detection start and end times corresponding to two equivalent detection volume units are read first. When two equivalent detection volume units correspond to the same construction cycle number, and both detection start and end times are within the detection period corresponding to that construction cycle, the construction stage consistency is determined to be at the first level. When two equivalent detection volume units correspond to adjacent construction cycle numbers, and the two detection start and end times are respectively within the detection period before the end of the previous construction cycle and the detection period after the start of the next construction cycle, the construction stage consistency is determined to be at the second level. When the construction cycle numbers corresponding to two equivalent detection volume units are different and not adjacent, or when the two detection start and end times span more than one construction cycle, the construction stage consistency is determined to be at the third level.

[0144] Once determined, the consistency of the construction phase is paired with the corresponding two equivalent detection volume units and recorded. This, along with mileage proximity, lateral overlap, azimuth consistency, and depth consistency, serves as the basis for forming the cross-source attribution score.

[0145] By obtaining cross-source attribution scores, the originally scattered mileage, lateral, directional, depth, and construction stage relationships among equivalent detection volume units from different detection methods can be unified into a single comparison result. On the one hand, cross-source attribution scores enable the subsequent establishment of candidate connection edges not only based on a single location proximity relationship, but also to consider whether spatial location, detection direction, detection depth, and construction stage correspond to each other, thereby reducing the possibility of incorrect connections between different fracture clusters. On the other hand, cross-source attribution scores provide a direct basis for subsequent elimination or retention of candidate connection edges, making the generation process of water-rich fracture candidate networks have a unified comparison basis and a unified retention basis.

[0146] Candidate connection edges are established based on cross-source attribution scores.

[0147] Methods for establishing candidate connection edges include:

[0148] For each equivalent detection volume unit in the multi-source anomaly volume table, cross-source pairing is performed according to the cross-source attribution score from high to low. Pairing relationships that originate from different detection methods, whose cross-source attribution score is in the first or second sorting position of all cross-source pairing results corresponding to the current equivalent detection volume unit, whose mileage proximity and lateral overlap do not belong to the third level, and whose orientation consistency, depth consistency, and construction stage consistency belong to the third level are determined as candidate connection edges; otherwise, the pairing relationship is eliminated.

[0149] For each equivalent detection volume unit in the multi-source anomaly volume table, cross-source pairing is performed according to the cross-source attribution score from high to low. Specifically, equivalent detection volume units are selected one by one as the current equivalent detection volume unit according to the record order in the multi-source anomaly volume table; then, the remaining equivalent detection volume units from different detection methods are extracted from the multi-source anomaly volume table, and the current equivalent detection volume unit is matched with the remaining equivalent detection volume units from different detection methods respectively; then, the cross-source attribution scores formed by each correspondence are read, and sorted according to the cross-source attribution scores from high to low to obtain the cross-source pairing order corresponding to the current equivalent detection volume unit; when the cross-source attribution scores corresponding to two cross-source pairs are the same, the cross-source pairing with more first-level numbers in mileage proximity and lateral overlap is given priority in the first sorting position; when the number of first-level numbers in mileage proximity and lateral overlap is the same, the cross-source pairing with more first-level numbers in azimuth consistency, depth consistency, and construction stage consistency is given priority in the first sorting position; when the number of first-level numbers is still the same, the sorting position is determined according to the record order in the multi-source anomaly volume table.

[0150] After sorting, the pairing relationship between the current equivalent probe volume unit and each paired equivalent probe volume unit is retained in order of sorting results as the basis for subsequent screening.

[0151] Pairing relationships originating from different detection methods, whose cross-source attribution scores are ranked in the first or second position of all cross-source pairing results corresponding to the current equivalent detection volume unit, whose mileage proximity and lateral overlap do not belong to the third level, and whose azimuth consistency, depth consistency, and construction stage consistency belong to the third level at most, are determined as follows: First, the cross-source attribution score ranking results are read according to all cross-source pairing results corresponding to the current equivalent detection volume unit; then, the pairing relationships corresponding to the first and second ranking positions are retained; subsequently, the mileage proximity, lateral overlap, azimuth consistency, depth consistency, and construction stage consistency corresponding to each pairing relationship are checked; when either mileage proximity or lateral overlap belongs to the third level, the corresponding pairing relationship is removed; when two or more of the azimuth consistency, depth consistency, and construction stage consistency belong to the third level simultaneously, the corresponding pairing relationship is removed; only pairing relationships that meet the candidate connection edge determination conditions are determined as candidate connection edges.

[0152] By establishing candidate connection edges, the numerous scattered cross-source pairings in the multi-source anomaly volume table can be further compressed into connection relationships that can still be used for network organization. On the one hand, candidate connection edges prioritize the retention of pairings with higher cross-source attribution scores that meet the retention criteria for candidate connection edges, so that the formation of subsequent water-rich fracture candidate networks no longer depends on all pairing results participating in the connection simultaneously, thereby reducing the situation where low-correspondence pairings are written into the network structure. On the other hand, candidate connection edges enable each equivalent detection volume unit to form a connection basis with equivalent detection volume units from different detection methods, thereby providing a direct data basis for subsequent elimination or retention of candidate connection edges and obtaining water-rich fracture candidate networks.

[0153] Candidate connection edges are either eliminated or retained to obtain a water-rich fracture candidate network.

[0154] Methods for obtaining candidate networks for water-rich fractures include:

[0155] For multiple candidate connection edges corresponding to the same equivalent detection volume unit, sort them according to the cross-source attribution score; determine the candidate connection edge that ranks first and comes from different detection methods as the priority connection edge; for candidate connection edges other than the priority connection edges, perform retention or rejection processing according to the corresponding sensitive direction relationship, cross-source attribution score, construction stage consistency and depth consistency of the candidate connection edge; form a water-rich fracture candidate network based on the retained candidate connection edges.

[0156] For multiple candidate connection edges corresponding to the same equivalent detection volume unit, they are sorted according to the cross-source attribution score. Specifically, all candidate connection edges corresponding to the same equivalent detection volume unit are selected one by one from the candidate connection edges; then, the cross-source attribution scores corresponding to all candidate connection edges are read, and sorted from high to low according to the cross-source attribution scores; when the cross-source attribution scores corresponding to multiple candidate connection edges are consistent, the candidate connection edges that come from different detection methods and have more first-level results in terms of mileage proximity, lateral overlap, azimuth consistency, depth consistency, and construction stage consistency are prioritized and placed in the first sorting position; when the first-level results are consistent, the candidate connection edges with the smallest mileage difference between phase correction coordinates and the largest lateral coverage overlap length are placed in the first sorting position; after sorting, the multiple candidate connection edges corresponding to the same equivalent detection volume unit are formed into an ordered sequence of candidate connection edges for subsequent elimination or retention processing.

[0157] For candidate connecting edges other than the priority connecting edges, retention or rejection is performed based on the corresponding sensitive direction correspondence, cross-source attribution score, construction stage consistency, and depth consistency. Specifically, firstly, candidate connecting edges ranked first and originating from different detection methods are determined as priority connecting edges; then, the sensitive direction correspondence between candidate connecting edges other than the priority connecting edges and the priority connecting edges is compared. The sensitive direction correspondence is determined by calculating the angle between the two sensitive directions in the coordinate system in front of the tunnel face; when the angle is not greater than a preset direction consistency threshold, the two sensitive directions are determined to point to the same forward coverage area; when the angle is greater than the preset direction consistency threshold, the two sensitive directions are determined not to point to the same area. A forward coverage area is defined as follows: When the sensitive direction corresponding to a candidate connecting edge other than the preferred connecting edge points to the same forward coverage area as the preferred connecting edge, the candidate connecting edge other than the preferred connecting edge is retained; when the sensitive direction corresponding to a candidate connecting edge other than the preferred connecting edge does not point to the same forward coverage area as the preferred connecting edge, and at least one of the consistency of construction stage and the consistency of depth is at level three, the candidate connecting edge other than the preferred connecting edge is removed; when the sensitive direction corresponding to a candidate connecting edge other than the preferred connecting edge does not point to the same forward coverage area as the preferred connecting edge, but neither the consistency of construction stage nor the consistency of depth is at level three, the candidate connecting edges other than the preferred connecting edge are retained in descending order of cross-source attribution score.

[0158] Based on the retained candidate connecting edges, a candidate network of water-rich fractures is formed. Specifically, all candidate connecting edges are first read in the order of retention; then, the continuity of adjacent candidate connecting edges is compared one by one; when two candidate connecting edges share a common equivalent detection volume unit, the common equivalent detection volume unit is determined as the receiving equivalent detection volume unit; when two candidate connecting edges do not share a common equivalent detection volume unit, the equivalent detection volume unit with the larger phase correction coordinate value on the face advancement direction axis is determined as the end equivalent detection volume unit of the previous candidate connecting edge, and the equivalent detection volume unit with the smaller coordinate value is determined as the starting equivalent detection volume unit of the next candidate connecting edge.

[0159] Subsequently, the phase correction coordinates and resolution radius corresponding to the equivalent detection volume unit at the end of the previous candidate connection edge and the equivalent detection volume unit at the beginning of the next candidate connection edge are read. The maximum value between the resolution radius corresponding to the equivalent detection volume unit at the end of the previous candidate connection edge and the resolution radius corresponding to the equivalent detection volume unit at the beginning of the next candidate connection edge is taken as the allowable value for cascading. When the mileage difference between the two phase correction coordinates does not exceed the allowable value for cascading, and the corresponding lateral overlap and depth consistency are not at level three, the two candidate connection edges are determined to have a sequential relationship. When the two candidate connection edges have a common equivalent detection volume unit, the phase correction coordinates corresponding to the common equivalent detection volume unit are directly used as the connection position of the sequential relationship. When there is a sequential relationship, and the detection methods corresponding to the adjacent candidate connection edges include at least two different detection methods, the adjacent candidate connection edges are assigned to the same water-rich fracture candidate network. When the above conditions are not met, the corresponding candidate connection edges are recorded as different water-rich fracture candidate networks.

[0160] The detection methods, phase correction coordinates, and cross-source attribution scores corresponding to each candidate connection edge in the water-rich fracture candidate network are associated and recorded for use in connectivity determination and risk classification. Specifically, the two equivalent detection volume units corresponding to each candidate connection edge are read one by one in the order of the candidate connection edges in the water-rich fracture candidate network. Then, the detection methods, phase correction coordinates, and cross-source attribution scores corresponding to the two equivalent detection volume units are extracted respectively. Subsequently, the detection methods, phase correction coordinates, and cross-source attribution scores corresponding to the same candidate connection edge are written into the same associated record. When the same water-rich fracture candidate network contains multiple candidate connection edges, they are written continuously in the connection order of the candidate connection edges in the water-rich fracture candidate network, thereby forming an associated record result that can be directly called for connectivity determination and risk classification.

[0161] Example 2:

[0162] like Figure 3 As shown, this embodiment provides a water-rich fracture network inversion system, including:

[0163] The parameter acquisition module is used to pre-establish a detection parameter file; and to collect the detection method, face advance record, spatial location of measuring point, detection start and end time, detection depth range and abnormal response data corresponding to each detection batch.

[0164] The phase correction module generates a multi-source original detection table based on the sensitive direction and resolution radius in the detection method and detection parameter file corresponding to each detection batch; based on the face advance record and the start and end times of detection, the multi-source original detection table is processed by advance phase backtracking to obtain phase correction coordinates unified with the target face position;

[0165] The volume equivalent module is used to construct an equivalent detection volume unit based on the sensitive direction, resolution radius, detection depth range and phase correction coordinates of each detection method in the multi-source original detection table.

[0166] The attribution scoring module is used to map the anomaly response data corresponding to the phase correction coordinates to the equivalent detection volume unit, and generate a multi-source anomaly volume table; the cross-source attribution score is calculated based on the multi-source anomaly volume table.

[0167] The network inversion module is used to establish candidate connection edges based on cross-source attribution scores; the candidate connection edges are then processed to eliminate or retain them to obtain a candidate network of water-rich fractures.

[0168] In summary, this application incorporates the target face position corresponding to the current inversion into the advance phase backtracking process, and uses the phase correction coordinates as the positioning reference for the equivalent detection volume unit. It further imposes deterministic constraints on the cross-source attribution score, the retention conditions of candidate connection edges, and the concatenation rules of the water-rich fracture candidate network. This enables the comparison of abnormal response data corresponding to different detection methods, different detection batches, and different construction stages under the same target face position, and completes cross-source pairing and network organization on a unified volume reference. This reduces the mis-splitting of the water-rich fracture candidate network caused by differences in advance position, inconsistent coordinate references, and uncertain pairing rules, and provides a basis for connectivity determination and risk classification.

[0169] The above description is merely a preferred embodiment of this application. The scope of protection of this application is not limited to the above embodiments. All technical solutions falling within the scope of this application's concept are within the scope of protection of this application. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of this application should also be considered within the scope of protection of this application.

Claims

1. A method for inverting water-rich fracture networks, characterized in that, include: Establish a pre-established detection parameter file; collect detection methods, face advance records, spatial locations of measuring points, detection start and end times, detection depth range, and abnormal response data for each detection batch; Based on the sensitive direction and resolution radius in the detection method and detection parameter file corresponding to each detection batch, a multi-source original detection table is generated; based on the face advance record and the start and end times of detection, the multi-source original detection table is subjected to advance phase backtracking processing to obtain phase correction coordinates unified with the target face position; Based on the sensitive direction, resolution radius, detection depth range and phase correction coordinates of each detection method in the original multi-source detection table, an equivalent detection volume unit is constructed; the abnormal response data corresponding to the phase correction coordinates is mapped to the equivalent detection volume unit to generate a multi-source anomaly volume table. Cross-source attribution scores were calculated based on the multi-source anomaly volume table; Candidate connection edges are established based on cross-source attribution scores; the candidate connection edges are then eliminated or retained to obtain a candidate network of water-rich fractures.

2. The water-rich fracture network inversion method according to claim 1, characterized in that, Methods for generating multi-source raw probe tables include: Based on the detection method corresponding to each detection batch, the corresponding sensitive direction and resolution radius are retrieved from the detection parameter file; Based on the tunnel face advancement records and the start and end times of the exploration, the tunnel face mileage and construction cycle number corresponding to each exploration batch are extracted; The detection method, working face mileage, construction cycle number, spatial location of measuring point, detection start and end time, detection depth range, abnormal response data, sensitive direction and resolution radius are written into the corresponding information to generate a multi-source original detection table.

3. The water-rich fracture network inversion method according to claim 1, characterized in that, Methods for obtaining phase correction coordinates uniform to the target face position include: Read the target face position corresponding to the current inversion; The advance positions of each probe batch at the start and end times of the probe were determined based on the advance records at the working face. Based on the mileage difference between the target face position and the advancing position, an advancing correction amount is generated; The spatial position of the measuring point is converted back along the advancing direction of the tunnel face according to the advancing correction amount to obtain the phase correction coordinates unified with the target tunnel face position.

4. The water-rich fracture network inversion method according to claim 1, characterized in that, Methods for constructing equivalent probe volume units include: Based on the sensitive directions corresponding to each detection method in the multi-source original detection table, the extension direction of the equivalent detection volume unit is determined; The lateral coverage range of the equivalent detection volume unit is determined based on the resolution radius; The longitudinal coverage range of the equivalent detection volume unit is determined based on the detection depth range; Using phase-corrected coordinates as the positioning reference for the equivalent detection volume unit, and combining the extension direction, lateral coverage range, and longitudinal coverage range corresponding to the same record in the multi-source original detection table, the equivalent detection volume unit is constructed.

5. The water-rich fracture network inversion method according to claim 4, characterized in that, Methods for generating multi-source anomaly volume tables include: Based on the phase correction coordinates and the anomaly response spatial interval corresponding to the anomaly response data, the equivalent detection volume unit corresponding to the anomaly response data is determined; When the phase correction coordinates fall within the overlapping range of multiple equivalent detection volume units, the target equivalent detection volume unit is determined according to the sensitive direction and detection depth range corresponding to the detection method. The equivalent detection volume unit of the target is written to correspond with the anomaly response data to generate a multi-source anomaly volume table.

6. The water-rich fracture network inversion method according to claim 4, characterized in that, Methods for obtaining cross-source attribution scores include: For the equivalent detection volume units in the multi-source anomaly volume table that originate from different detection methods, the mileage proximity, lateral overlap, azimuth consistency, depth consistency, and construction stage consistency are calculated respectively. A cross-source attribution score is formed based on mileage proximity, lateral overlap, orientation consistency, depth consistency, and construction stage consistency.

7. The water-rich fracture network inversion method according to claim 6, characterized in that, Methods for calculating mileage proximity, lateral overlap, azimuth consistency, depth consistency, and construction stage consistency include: Based on the phase correction coordinates, the mileage difference and lateral overlap relationship between equivalent detection volume units from different detection methods are extracted to form mileage proximity and lateral overlap. Based on the correspondence between sensitive directions and detection depth ranges, azimuth consistency and depth consistency are established; Based on the correspondence between the construction cycle number and the start and end times of the detection, a consistency of construction stages is established; Mileage proximity, lateral overlap, orientation consistency, depth consistency, and construction stage consistency all include first, second, and third levels.

8. The water-rich fracture network inversion method according to claim 7, characterized in that, Methods for establishing candidate connection edges include: For each equivalent detection volume unit in the multi-source anomaly volume table, cross-source pairing is performed according to the cross-source attribution score from high to low. Pairing relationships originating from different detection methods, whose cross-source attribution scores are located in the first or second sorting position of all cross-source pairing results corresponding to the current equivalent detection volume unit, whose mileage proximity and lateral overlap do not belong to the third level, and whose azimuth consistency, depth consistency, and construction stage consistency belong to the third level, are identified as candidate connection edges; otherwise, the pairing relationships are eliminated.

9. The water-rich fracture network inversion method according to claim 8, characterized in that, Methods for obtaining candidate networks for water-rich fractures include: For multiple candidate connection edges corresponding to the same equivalent detection volume unit, sort them according to the cross-source attribution score; Candidate connecting edges that are ranked first and originate from different detection methods are determined as priority connecting edges; For candidate connection edges other than the priority connection edges, retain or remove them based on the sensitive direction correspondence, cross-source attribution score, construction stage consistency and depth consistency of the candidate connection edges; A candidate network of water-rich fractures is formed based on the retained candidate connection edges.

10. The water-rich fracture network inversion method according to claim 9, characterized in that, The detection method, phase correction coordinates, and cross-source attribution score corresponding to each candidate connection edge in the water-rich fracture candidate network are associated and recorded for use in connectivity determination and risk classification.

11. A water-rich fracture network inversion system, used to implement the water-rich fracture network inversion method according to any one of claims 1-10, characterized in that, include: The parameter acquisition module is used to pre-establish a detection parameter file; and to collect the detection method, face advance record, spatial location of measuring point, detection start and end time, detection depth range and abnormal response data corresponding to each detection batch. The phase correction module generates a multi-source original detection table based on the sensitive direction and resolution radius in the detection method and detection parameter file corresponding to each detection batch; based on the face advance record and the start and end times of detection, the multi-source original detection table is processed by advance phase backtracking to obtain phase correction coordinates unified with the target face position; The volume equivalent module is used to construct an equivalent detection volume unit based on the sensitive direction, resolution radius, detection depth range and phase correction coordinates of each detection method in the multi-source original detection table. The attribution scoring module is used to map the abnormal response data corresponding to the phase correction coordinates to the equivalent detection volume unit and generate a multi-source anomaly volume table. Cross-source attribution scores were calculated based on the multi-source anomaly volume table; The network inversion module is used to establish candidate connection edges based on cross-source attribution scores; the candidate connection edges are then processed to eliminate or retain them to obtain a candidate network of water-rich fractures.

Images