Metro structure safety related geological defect geophysical prospecting surveying and mapping cooperative inspection method and system

Abstract

The invention discloses a geophysical prospecting, surveying and mapping collaborative inspection method and system for safety-related geological defects of a subway structure, and particularly relates to the technical field of subway engineering safety detection, and the method comprises the five steps of early-stage preprocessing and detection area division, multi-source collaborative detection, detection data fusion and mapping, collaborative inspection, and dynamic updating and operation and maintenance adaptation. Gravity exploration is taken as a core, multiple detection means are fused, the weight of gravity data is enhanced in combination with an improved BP neural network model, the result is ensured to be reliable through a triple inspection mechanism, and a whole-process closed-loop process is constructed; the system comprises a main control module and five functional modules, and is accurately matched with the steps of the method, so that integrated cooperative operation is realized. The method improves the detection precision of geological defects and earthquake susceptible structures, reduces the manual error and operation and maintenance cost, is suitable for the full-stage application of a subway, provides precise data support for the safety prevention and control of a subway structure, and is remarkable in practicability and application value.

Description

Technical Field

[0001] This invention relates to the field of subway engineering safety inspection technology, and more specifically, to a collaborative inspection method and system for detecting geological defects related to subway structural safety. Background Technology

[0002] The geological conditions along the subway line are complex. Geological defects such as shallow karst cavities and abnormal groundwater bodies, as well as earthquake-prone geological structures such as deep active fault zones and sand liquefaction layers, all pose long-term potential threats to the safety of the subway structure. Especially during the subway operation and maintenance phase, the evolution of geological defects and changes in the activity of geological structures can easily lead to safety hazards such as tunnel settlement and lining cracking. Therefore, accurate detection, mapping and collaborative inspection of geological defects and earthquake-prone geological structures are of paramount importance.

[0003] Currently, most subway geological defect detection methods rely on single geophysical detection techniques, which suffer from insufficient detection accuracy and a disconnect between shallow and deep geological information. Furthermore, the detection, mapping, and verification processes are independent of each other and lack a collaborative verification mechanism, which can easily lead to significant deviations between the detection results and the actual geological conditions. In addition, existing detection technologies often lack deep integration with subway operation and maintenance management, and there is no dynamic update mechanism, making it difficult to track the evolution of geological defects in real time and providing accurate and continuous data support for subway structural safety control.

[0004] In addition, existing detection systems mostly only realize a single detection function, have poor adaptability to detection methods, and are difficult to complete the entire process of detection-map-inspection-feedback. There are many manual intervention links, which not only result in low inspection efficiency, but also easily introduce human error. They cannot meet the high-precision, high-efficiency detection, mapping and collaborative inspection needs of subway engineering survey, construction and operation and maintenance.

[0005] Therefore, there is an urgent need to develop a collaborative inspection method and system for detecting and mapping geological defects in subways that integrates multiple methods, maps and inspects, and is adapted to operation and maintenance needs, in order to address the shortcomings of existing technologies. Summary of the Invention

[0006] To overcome the aforementioned deficiencies of the prior art, this invention provides a collaborative inspection method and system for detecting and mapping geological defects related to subway structural safety, in order to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] On the one hand, this invention provides a collaborative inspection method for detecting and mapping geological defects related to subway structural safety, including the following steps:

[0009] S1. Collect basic data or information along the subway line, delineate the core detection area and extended detection area, conduct detection environment calibration and establish an error correction model;

[0010] S2. Taking gravity exploration as the core, integrating micro-motion detection, ground-penetrating radar detection, and elastic wave CT detection, collect and preprocess multi-source detection data to initially identify geological defects and earthquake-prone geological structures;

[0011] S3. Integrate the preprocessed detection data through a multi-source data fusion model, draw a distribution map of layered geological defects and earthquake-prone geological structures, and perform an initial verification of the drawing accuracy.

[0012] S4. Through a triple inspection mechanism, establish the correlation between geological defects, earthquake-prone geological structures and subway structural safety, evaluate the inspection results and generate feedback.

[0013] S5. Regularly conduct detection and verification, update the mapping results and 3D model, connect the collaborative inspection results with the subway operation and maintenance management system, output targeted operation and maintenance suggestions, and then use the updated results as input to feed back into the error correction model in step S1 and the multi-source collaborative detection scheme in step S2, so as to realize the iterative tracking of the dynamic evolution of geological defects and the adaptive optimization of the geological model.

[0014] Preferably, in step S1, the basic data or information includes geological survey reports along the subway line, subway structural design parameters, existing geophysical exploration data, and regional seismic geological data;

[0015] The core detection area is within 50m around the subway tunnel, used to investigate shallow geological defects;

[0016] The extended detection area is within 100-200m outside the core detection area, used to investigate deep earthquake-prone geological structures;

[0017] The method for constructing and using the error correction model is as follows: collect historical data, including basic geological parameter samples and their corresponding multi-source detection error data. Use the basic geological parameter samples as input and the corresponding detection error data as the expected output to conduct supervised training on the BP neural network, so that the network learns the mapping relationship from geological parameters to errors. The trained model can automatically predict and output the corresponding error correction parameters for new basic geological parameters.

[0018] Preferably, in step S3, the multi-source data fusion model is an improved BP neural network fusion model, the improvement being: assigning higher initial connection weights to the gravity exploration data in the input layer, and applying normalization processing to the gravity data independent of other data sources.

[0019] Draw a distribution map of layered geological defects and earthquake-prone geological structures, specifically including: generating a distribution map of shallow geological defects, a distribution map of deep earthquake-prone geological structures, and a three-dimensional comprehensive distribution map. The initial verification error of the drawing accuracy should be ≤5%.

[0020] Preferably, in step S4, the triple verification mechanism includes detection-drawing comparison verification, drawing-structural safety correlation verification, and on-site practical verification. The pass criterion for detection-drawing comparison verification is that the consistency between the two is ≥95%, and the pass criterion for drawing-structural safety correlation verification is that the prediction error of its correlation model is ≤5%.

[0021] In step S5, the dynamic update cycle is every 6-12 months, and gravity exploration and micro-motion detection are carried out regularly for verification.

[0022] On the other hand, the present invention provides a collaborative inspection system for detecting and mapping geological defects related to subway structural safety, used to implement the above-mentioned method, including:

[0023] The main control module, and the components electrically connected to and controlled by the main control module:

[0024] The preprocessing and area division module is used to perform step S1, including collecting basic data or information along the subway line, delineating the core detection area and the extended detection area, and establishing an error correction model based on a BP neural network.

[0025] The multi-source collaborative detection module is used to perform step S2, which includes data acquisition based on gravity exploration and integrating micro-motion detection, ground-penetrating radar detection and elastic wave CT detection, as well as preprocessing the acquired multi-source detection data.

[0026] The data fusion and plotting module is used to perform step S3, including fusing the preprocessed multi-source detection data using an improved BP neural network fusion model, and plotting shallow geological defect distribution maps, deep earthquake-prone geological structure distribution maps, and three-dimensional comprehensive distribution maps based on the fusion results.

[0027] The collaborative inspection module is used to execute step S4, including conducting detection-drawing comparison inspection, drawing-structural safety correlation inspection, and on-site practical inspection, and establishing correlation relationships and generating feedback based on the inspection results;

[0028] The dynamic update and operation and maintenance adaptation module is used to execute step S5, including triggering detection and verification according to a preset cycle to update the drawing and model, and connecting the results to the subway operation and maintenance management system.

[0029] Preferably, the multi-source collaborative detection module includes a gravity exploration unit, a micro-motion detection unit, a ground-penetrating radar detection unit, an elastic wave CT detection unit, and a data preprocessing unit;

[0030] The gravity exploration unit is configured to use a high-precision gravimeter and to set up detection points in a grid pattern to collect gravity anomaly data and gravity gradient data.

[0031] The data preprocessing unit is configured to perform noise reduction, correction, and standardization on the data from each detection unit.

[0032] Preferably, the correction operations performed by the data preprocessing unit on the data collected by the gravity exploration unit specifically include terrain correction, Bouguer correction, and equilibrium correction.

[0033] Preferably, the data fusion and plotting module includes:

[0034] The data fusion unit is configured to run an improved BP neural network fusion model, which integrates multi-source data by assigning higher initial connection weights to gravity exploration data and performing independent normalization processing, and outputs the fused geological body parameters.

[0035] The drawing unit is configured to generate shallow geological defect distribution maps, deep earthquake-prone geological structure distribution maps, and three-dimensional comprehensive distribution maps based on the geological body parameters output by the data fusion unit.

[0036] Preferably, the collaborative verification module is configured as follows:

[0037] The plotting results output by the data fusion and plotting module are compared with the original detection data from the multi-source collaborative detection module to calculate consistency.

[0038] By combining the subway structural design parameters, the associated model is called to calculate the risk probability of the geological structures marked in the drawing results to the safety of the subway structure;

[0039] It receives input from on-site practical tests, comprehensively evaluates the test results, and generates feedback information including operation and maintenance suggestions.

[0040] Preferably, the dynamic update and operation and maintenance adaptation module is configured as follows:

[0041] Commands are sent to the multi-source collaborative detection module every 6 to 12 months to initiate gravity exploration and micro-motion detection verification.

[0042] Update the drawing results and 3D model in the data fusion and drawing module according to the verification data;

[0043] The updated model, test results, and operation and maintenance suggestions will be pushed to the subway operation and maintenance management system.

[0044] The updated drawing results are then fed back to the error correction model of the preprocessing and region division module and the detection scheme configuration of the multi-source collaborative detection module.

[0045] The technical effects and advantages of this invention are as follows:

[0046] This invention effectively solves the technical problems of disconnect between detection, mapping, and inspection in existing subway geological defect detection technologies, as well as insufficient accuracy in identifying earthquake-prone geological structures. By constructing a multi-source collaborative detection mode with gravity exploration as the core, integrating micro-motion detection, ground-penetrating radar detection, and elastic wave CT detection, and combining an improved BP neural network fusion model to enhance the weight of gravity data, it achieves accurate identification of shallow geological defects and deep earthquake-prone geological structures. At the same time, a triple inspection mechanism ensures the authenticity and reliability of the detection and mapping results. Compared with traditional single detection methods, the detection accuracy is improved by more than 50%, and the mapping accuracy error is controlled within 5%. This significantly improves the efficiency and accuracy of collaborative inspection of subway geological defect detection and mapping, providing accurate data support for subway structural safety control.

[0047] This invention achieves deep collaborative adaptation between the method and the system. Through five pre-set functional modules, including a preprocessing and area division module and a multi-source collaborative detection module, each step of the detection and mapping collaborative inspection method is corresponding to the other. This enables a closed-loop operation of the entire process of detection-mapping-inspection-feedback-optimization, which can complete the entire inspection process without much human intervention. This effectively reduces the intensity of manual operation and human error, and greatly improves inspection efficiency. At the same time, the system has human-computer interaction, data storage and export functions, and can retain all process data for easy subsequent query, traceability and archiving. It is highly practical and adaptable to the actual needs of long-term operation and maintenance of subway projects.

[0048] This invention boasts wide adaptability and strong operation and maintenance orientation, making it flexibly applicable to multiple stages of subway surveying, construction, and operation and maintenance. By adjusting parameters such as the spacing between detection points and the dynamic update cycle, it can adapt to actual scenarios with different geological conditions and subway structural parameters. Simultaneously, through a dynamic update mechanism, it regularly conducts detection verification, tracks the evolution trend of geological defects and changes in seismic susceptibility risks in real time, and deeply integrates the inspection results with the subway operation and maintenance management system to output targeted operation and maintenance suggestions. This achieves precise adaptation between the detection and mapping inspection results and subway operation and maintenance work, effectively reducing potential safety hazards in subway structures, extending the service life of subway structures, and reducing operation and maintenance costs. It has significant economic and social value and broad application prospects. Attached Figure Description

[0049] Figure 1 This is a flowchart of the overall method of the present invention.

[0050] Figure 2 This is a system diagram of the present invention.

[0051] The attached diagram is labeled as follows: 1. Main control module; 2. Preprocessing and area division module; 3. Multi-source collaborative detection module; 300. Gravity exploration unit; 301. Micromotion detection unit; 302. Ground penetrating radar detection unit; 303. Elastic wave CT detection unit; 304. Data preprocessing unit; 4. Detection map fusion and mapping module; 400. Data fusion unit; 401. Layered mapping unit; 5. Collaborative verification module; 6. Dynamic update and operation and maintenance adaptation module. Detailed Implementation

[0052] The present invention will be further described in detail below with reference to specific embodiments, in order to make the technical solution, implementation process and beneficial effects of the present invention clearer and easier to understand, to ensure that the present invention is fully disclosed, and to facilitate understanding and implementation by those skilled in the art.

[0053] It should be noted that the following embodiments are only used to explain the present invention and are not intended to limit the scope of protection of the present invention. Any non-substantial modifications, equivalent substitutions, or improvements made by those skilled in the art based on the present invention should fall within the scope of protection of the present invention. In addition, the specific technical parameters, equipment models, etc. mentioned in the following embodiments are only examples adapted to actual implementation scenarios. Those skilled in the art can flexibly adjust them according to actual needs, as long as the technical effects defined in the claims of the present invention can be achieved.

[0054] Example 1

[0055] As attached Figure 1 As shown in the figure, this invention provides a collaborative inspection method for detecting and mapping geological defects related to subway structural safety. This method is applied during the operation and maintenance phase of subway projects. The subway line traverses urban-rural fringe areas, with shallow karst cavities and abnormal groundwater zones along its route. It also passes through extended sections of regional fault zones, containing deep active fault zones and liquefiable sand layers, among other earthquake-prone geological structures. During long-term operation and maintenance, it is crucial to accurately grasp the evolution patterns of geological defects and the risk of earthquake susceptibility. The method and system of this invention enable efficient and accurate collaborative inspection and mapping, ensuring the safety of the subway structure. The specific implementation process is as follows:

[0056] S1. Pre-processing and detection area delineation

[0057] First, basic data or information along the subway project route should be collected, specifically including: the geological survey report of the entire line (including parameters such as soil and rock type, groundwater level, and karst development), subway tunnel structural design parameters (tunnel diameter 6.0m, burial depth 12-28m, lining using C50 concrete), existing InSAR surface deformation data, and regional seismic geological data (this area is a peak ground acceleration zone of 0.10g, with an extension of a regional active fault zone).

[0058] Based on the above basic data or information, a core detection area and an extended detection area are delineated: the core detection area is within 50m of the subway tunnel, focusing on investigating shallow geological defects (karst cavities, abnormal groundwater areas, voids behind the lining, etc.); the extended detection area is within 100-200m outside the core detection area (150m in this example), focusing on investigating deep earthquake-prone geological structures (active fault zones, liquefaction layers of sand, etc.).

[0059] Subsequently, the detection environment was calibrated: ground electromagnetic interference sources in the detection area, such as high-voltage lines, large metal components, and areas surrounding communication base stations, were removed; the gravity exploration points were leveled to avoid ground undulations and debris accumulation affecting the accuracy of gravity measurements; at the same time, an error correction model was constructed based on the BP neural network algorithm, linking basic geological parameters (soil type, groundwater level) with the detection error, and preset error correction parameters were used for error correction of subsequent multi-source detection data to ensure the accuracy of the detection data.

[0060] S2. Multi-source collaborative detection: Acquiring and preprocessing multi-source detection data.

[0061] This step focuses on gravity exploration, integrating micro-motion detection, ground-penetrating radar detection, and elastic wave CT detection. The specific implementation process is as follows:

[0062] S2.1 Gravity Exploration: A CG-5 high-precision gravimeter (measurement accuracy ±0.01mGal) was used to deploy grid-like detection points in the core and extended detection areas. The spacing between points in the core detection area was 8m, and the spacing between points in the extended detection area was 15m. Gravity anomaly data and gravity gradient data were collected simultaneously. Through methods such as regional-residual anomaly separation, three-dimensional Eulerian deconvolution, and three-dimensional gravity inversion, the density distribution of underground geological bodies was inverted to preliminarily identify density anomaly areas. Low-density areas were preliminarily identified as karst cavities and active fault zones, while high-density areas were identified as hard rock masses, providing basic data for subsequent identification of earthquake-prone geological structures.

[0063] S2.2 Micromotion Detection: A circular micromotion detection array (array radius 5m, containing 7 detection nodes) is deployed in the above-mentioned gravity anomaly area to collect Rayleigh wave dispersion curves and invert the wave velocity structure in the 20-100 meter depth range underground. No artificial seismic source is required, which is suitable for the complex environment of urban operation and maintenance. It can supplement the identification of deep earthquake-prone geological structures (wave velocity anomaly areas corresponding to active fault zones and wave velocity reduction areas corresponding to sand liquefaction layers) and verify the authenticity of the density anomaly areas identified by gravity exploration.

[0064] S2.3 Ground-penetrating radar detection: A ground-penetrating radar antenna in the 1GHz band is used to lay detection lines along the subway line axis and both sides, with a detection line spacing of 5m. The focus is on detecting shallow geological defects in the 0.5-30m depth. Ground-penetrating radar reflected wave signals are collected, and the location, size, and morphological parameters of shallow geological defects are obtained through signal analysis. Three karst cavities and one groundwater anomaly zone are identified.

[0065] S2.4 Elastic wave CT detection: In the overlapping area of ​​the gravity anomaly zone and the micro-motion detection anomaly zone, i.e. the extension section of the preliminarily determined active fault zone, the excitation point and the receiving point are set up using the existing maintenance well holes of the subway. The well hole spacing is 25m. Elastic wave signals are emitted and reflected and transmitted signals are received. Through signal processing, elastic wave CT images are generated to accurately identify the width of the active fault zone and the nature of the filling material (in this embodiment, the active fault zone is about 4m wide and the filling material is loose sand), and to verify the activity of the earthquake-prone geological structure.

[0066] S2.5 Data Preprocessing: The above-mentioned multi-source detection data are uniformly preprocessed, specifically including: performing terrain correction, Bouguer correction, and equalization correction on gravity data to eliminate the influence of terrain undulation and shallow soil inhomogeneity; performing spectrum analysis and filtering on micromotion detection data to eliminate interference from urban environmental noise (such as noise generated by vehicle traffic and human activities); performing amplitude correction and phase correction on ground-penetrating radar data to eliminate the influence of reflected wave distortion; performing waveform inversion and error correction on elastic wave CT data to improve image resolution; finally, using a data standardization algorithm, the multi-source detection data are unified to the WGS84 coordinate system and the same numerical range to eliminate data fusion conflicts and provide a standardized data source for subsequent data fusion.

[0067] S3. Data fusion and geological defect mapping are performed, and the mapping accuracy is initially verified.

[0068] This step employs an improved BP neural network fusion model to integrate preprocessed multi-source detection data. The specific implementation process is as follows:

[0069] S3.1 Multi-source data fusion: Preprocessed gravity data, micro-motion detection data, ground-penetrating radar data, and elastic wave CT data are used as inputs to the improved BP neural network fusion model. The actual geological data (rock and soil density, wave velocity, and other parameters obtained from borehole sampling) in the geological survey report are used as labels to train the fusion model (100 iterations, convergence error ≤ 0.001). During the fusion process, the weight of gravity data is strengthened (weight ratio 40%) to ensure the accuracy of identifying earthquake-prone geological structures. Finally, the fused geological body parameters (density, wave velocity, resistivity) are output.

[0070] S3.2 Layered Mapping: Based on the fused geological body parameters, three-dimensional modeling and visualization techniques are used, combined with a GIS spatiotemporal database, to create three types of layered distribution maps, as follows:

[0071] Shallow geological defect distribution map (0-30m): Mark the location, size and shape of 3 karst cavities and 1 groundwater anomaly area, associate them with the lining mileage of the subway tunnel, and clarify the spatial relationship between the defects and the subway tunnel (e.g., karst cavity No. 1 is located 10m to the left of the tunnel, buried at a depth of 16m, with dimensions of 4m×3m×2m).

[0072] Distribution map of deep earthquake-prone geological structures (30-100m): Mark the distribution range, strike and dip of active fault zones, and mark the earthquake risk level in combination with regional seismic geological data (the extension of active fault zones is a medium earthquake risk zone, and the sandy soil liquefaction layer is a low earthquake risk zone).

[0073] Three-dimensional integrated distribution map: Integrating shallow and deep data, a three-dimensional digital twin model of underground geological bodies along the subway line is constructed, which intuitively presents the spatial relationship between geological defects, earthquake-prone geological structures and subway tunnels, and marks the hazard level of each defect and structure (karst cavities #1 and #2 are of medium hazard level, karst cavity #3 is of low hazard level, and active fault zone is of medium hazard level).

[0074] S3.3 Initial verification of mapping accuracy: Using on-site borehole sampling and UAV-borne ground-penetrating radar verification, 12% of the detection points are randomly selected (36 points are selected in this embodiment). The actual geological data of the boreholes are compared with the mapping data to calculate the mapping accuracy error. In this embodiment, the mapping accuracy error is 3.5%. If the error is ≤5%, the mapping result is qualified and proceeds to the next step. If the error is greater than 5%, return to step S2 and repeat the multi-source detection and data fusion until the accuracy meets the standard.

[0075] S4. Collaborative testing: Establishing correlations and evaluating test results.

[0076] This step establishes the correlation between geological defects, earthquake-prone geological structures, and subway structural safety through a triple verification mechanism (detection-map comparison verification, map-structural safety correlation verification, and on-site practical verification). The specific implementation process is as follows:

[0077] S4.1 Detection-Mapping Comparison Verification: The multi-source detection raw data (gravity anomaly data, micro-motion wave velocity data, ground-penetrating radar reflection wave data, etc.) collected in step S2 are compared with the layered distribution map and three-dimensional comprehensive distribution map drawn in step S3 to verify their consistency. In this embodiment, the consistency between the detection data and the mapping results is 96.8%, ≥95%, with no obvious deviation, and the mapping results are true and reliable.

[0078] S4.2, Drawing-Structural Safety Correlation Verification: Combining the subway tunnel structural design parameters (lining strength, stiffness) and real-time settlement monitoring data (using BOTDR fiber optic sensing technology, monitoring tunnel settlement ≤0.15mm / d), a correlation model was established between geological defects, earthquake-prone geological structures, and subway structural safety. Probabilistic seismic hazard analysis was used to calculate the probability of subway structural risk under different earthquake return periods. The verification results show that the minimum distance between the three karst cavities and the tunnel lining is ≥8m, which will not cause safety hazards such as lining cracking or leakage; under seismic action (50-year return period PGA=0.10g), the predicted tunnel settlement of the active fault zone is 0.5mm / d, which does not exceed the safety limit and is classified as medium risk; the sand liquefaction layer has no significant impact on the subway structure.

[0079] S4.3 On-site practical verification: Eight on-site verification points were set up in the medium-risk defect areas (karst cavities #1 and #2) and the medium-seismic-risk areas (extended sections of active fault zones) marked in the drawing results. Methods such as borehole CT and ultrasonic detection were used to verify the authenticity, size, and activity of earthquake-prone structures in the field. The on-site verification results were consistent with the drawing results and the prediction results of the correlation model. Simultaneously, on-site monitoring of the stress and settlement data of the subway tunnel lining was conducted, and the prediction results of the correlation model were compared. The prediction error of the correlation model was 2.9%, ≤5%, verifying the accuracy and reliability of the model.

[0080] S4.4 Evaluation and Feedback of Inspection Results: Based on the above triple inspection results, the collaborative inspection of this detection and mapping is deemed qualified (detection-mapping). Figure 1 With a consistency of ≥95% and a correlation model prediction error of ≤5%, the final output includes a layered geological defect distribution map, a deep earthquake-prone geological structure distribution map, a three-dimensional comprehensive distribution map, and a collaborative inspection report. For karst cavities of medium risk level and active fault zones of medium earthquake risk, preliminary operation and maintenance feedback suggestions are generated to provide a basis for subsequent dynamic updates and operation and maintenance adaptation.

[0081] S5, dynamic updates and operation and maintenance adaptation form a closed-loop process.

[0082] This step establishes a dynamic update mechanism to achieve deep adaptation between the detection and inspection results and subway operation and maintenance:

[0083] Dynamic updates: The dynamic update cycle is set at 8 months. Gravity exploration and micro-motion detection are carried out regularly along the subway project line to supplement and collect data on changes in geological defects (karst cavities) and earthquake-prone geological structures (active fault zones). The three-dimensional digital twin model and various drawing results are updated to track the evolution trend of defects and changes in earthquake susceptibility risk in real time.

[0084] Operation and Maintenance Adaptation: The results of this collaborative inspection and subsequent dynamic updates will be integrated with the metro operation and maintenance management system to output targeted operation and maintenance recommendations: For the karst cavities #1 and #2, which are of medium risk, it is recommended to conduct targeted monitoring every 6 months; for the extension of the active fault zone with medium seismic risk, it is recommended to conduct gravity exploration verification every 4 months to optimize the monitoring frequency; for the karst cavities #3 and the sand liquefaction layer, which are of low risk, it is recommended to conduct routine verification every 12 months. At the same time, the inspection results will be incorporated into the metro operation and maintenance archives to provide precise support for the safe operation and maintenance of the metro structure, forming a closed-loop process of detection-drawing-inspection-feedback-optimization.

[0085] Example 2

[0086] As attached Figure 2 As shown, this embodiment, based on Embodiment 1, provides a collaborative inspection system for detecting and mapping geological defects related to subway structural safety. Specifically, it includes a main control module 1, and a preprocessing and region division module 2, a multi-source collaborative detection module 3, a detection mapping fusion and drawing module 4, a collaborative inspection module 5, and a dynamic update and maintenance adaptation module 6, all electrically connected to the main control module 1. Each module works collaboratively, and the specific structure and functions are implemented as follows:

[0087] The preprocessing and area division module 2 performs all operations in step S1. Its specific functions include: collecting basic data or information along the subway line (geological survey reports, structural design parameters, existing geophysical exploration data, regional seismic and geological data); standardizing and classifying the input of various types of basic data or information through a built-in data integration unit; delineating core and extended exploration areas based on the basic data or information; outputting exploration area division maps and parameters; conducting exploration environment calibration, eliminating environmental interference factors, and leveling exploration points; and constructing an error correction model based on a BP neural network algorithm, outputting error correction parameters to support error correction for subsequent multi-source exploration data. This module is electrically connected to the main control module 1, transmitting the exploration area division results and error correction parameters to the main control module 1 for use by subsequent modules.

[0088] The multi-source collaborative detection module 3, which is the core of the system, is used to execute the operation in step S2. It includes a gravity exploration unit 300, a micro-motion detection unit 301, a ground-penetrating radar detection unit 302, an elastic wave CT detection unit 303, and a data preprocessing unit 304. The specific implementation is as follows:

[0089] Gravity exploration unit 300: Employs a CG-5 high-precision gravimeter, automatically deploys grid-like detection points based on the detection area parameters output by the preprocessing and regional division module 2, and collects gravity anomaly data and gravity gradient data; It has a built-in data processing subunit that uses methods such as regional-residual anomaly separation, three-dimensional Eulerian deconvolution, and three-dimensional gravity inversion to invert the density distribution of underground geological bodies, preliminarily identify density anomaly areas and preliminary information on earthquake-prone geological structures, and transmits the collected raw data and preliminary identification results to the data preprocessing unit 304 and the main control module 1;

[0090] Micromotion detection unit 301: includes a micromotion detection array and a spectrum analysis unit. Based on the gravity anomaly parameters output by gravity exploration unit 300, a circular micromotion detection array is deployed in the gravity anomaly area to collect Rayleigh wave dispersion curves. The spectrum analysis unit processes the collected signals to invert the wave velocity structure within a depth range of 20-100 meters underground, supplement the identification of deep earthquake-prone geological structures, verify the gravity exploration results, and transmit the micromotion detection data and verification results to the data preprocessing unit 304.

[0091] Ground-penetrating radar detection unit 302: Using a 1GHz frequency band ground-penetrating radar antenna, a detection line is laid out along the subway line axis and both sides to collect reflected wave signals of shallow geological defects, analyze and obtain the location, size and morphological parameters of the defects, and transmit the ground-penetrating radar detection data to the data preprocessing unit 304.

[0092] Elastic wave CT detection unit 303: includes an excitation device, a receiving device and a CT imaging unit. Based on the parameters of the overlapping area between the gravity anomaly zone and the micromotion detection anomaly zone, it uses the existing well holes of the subway to set up excitation and receiving points, transmits and receives elastic wave signals, generates elastic wave CT images through the CT imaging unit, accurately identifies fault fracture zone parameters, verifies the activity of earthquake-prone geological structures, and transmits elastic wave CT data and verification results to data preprocessing unit 304.

[0093] Data preprocessing unit 304: Receives multi-source detection data transmitted from the above four detection units, performs targeted noise reduction and correction processing on gravity data, micro-motion detection data, ground-penetrating radar data, and elastic wave CT data respectively, and then uses a data standardization algorithm to unify the multi-source detection data to the same coordinate system and numerical range to eliminate data fusion conflicts; transmits the preprocessed standardized detection data to the detection mapping fusion and drawing module 4, and simultaneously saves it back to the data storage unit of the main control module 1.

[0094] The detection, mapping, fusion, and plotting module 4 is used to perform the operation in step S3. It includes a data fusion unit 400 and a layered plotting unit 401, and is implemented as follows:

[0095] Data fusion unit 400: Employs an improved BP neural network fusion model, receives standardized detection data transmitted by data preprocessing unit 304, and uses actual geological data from the geological exploration report transmitted by preprocessing and regional division module 2 as labels to train the fusion model. Through iterative optimization of the model, it achieves deep fusion of multi-source data and outputs the fused geological body parameters (density, wave velocity, resistivity). During the fusion process, the weight of gravity data is strengthened to improve the identification accuracy of earthquake-prone geological structures. The fused geological body parameters are then transmitted to the layered mapping unit 401 and the main control module 1.

[0096] Layered drawing unit 401: Receives fused geological body parameters transmitted by data fusion unit 400, combines GIS spatiotemporal database and 3D modeling technology to draw shallow geological defect distribution maps, deep seismically susceptible geological structure distribution maps, and 3D comprehensive distribution maps, marking the parameters, hazard levels, and spatial relationships with subway structures of defects and seismically susceptible structures; Built-in drawing accuracy initial verification subunit: Using on-site borehole sampling and UAV-borne ground-penetrating radar verification, 10%-15% of detection points are randomly selected, and the actual geological data is compared with the drawing data to calculate the accuracy error. If the error is ≤5%, the drawing result is output to collaborative verification module 5; if the error is >5%, a feedback signal is sent to main control module 1, which instructs multi-source collaborative detection module 3 and data fusion unit 400 to re-perform the relevant operations until the drawing accuracy meets the standard.

[0097] Collaborative inspection module 5 is used to execute the operation in step S4. Through a triple inspection mechanism, it establishes the correlation between geological defects, earthquake-prone geological structures, and subway structural safety. The specific implementation is as follows:

[0098] The system receives the mapping results output from the detection-mapping fusion and mapping module 4, and simultaneously receives the raw detection data output from the multi-source collaborative detection module 3 and the basic data or information output from the preprocessing and area division module 2. It then conducts a triple verification: detection-mapping comparison verification (comparing the consistency between the raw detection data and the mapping results), mapping-structural safety correlation verification (establishing a correlation model based on subway structural design parameters and real-time settlement monitoring data to calculate the structural risk probability), and on-site practical verification (setting up verification points in high / medium-risk areas to verify defects and structural parameters on-site). The verification results are then evaluated. If the detection-mapping... Figure 1 If the consistency is ≥95% and the correlation model prediction error is ≤5%, the inspection is deemed qualified, and the final detection distribution map and collaborative inspection report are output, forming operation and maintenance feedback suggestions, which are transmitted to the dynamic update and operation and maintenance adaptation module 6 and the main control module 1; if the inspection fails, a feedback signal is sent to the corresponding module, instructing the relevant operations to be carried out again.

[0099] Dynamic update and operation and maintenance adaptation module 6, this module is used to perform the operation in step S5, and the specific implementation is as follows:

[0100] The system receives the inspection report and maintenance feedback suggestions output by the collaborative inspection module 5, establishes a dynamic update mechanism, and instructs the multi-source collaborative detection module 3 to conduct gravity exploration and micro-motion detection verification every 6-12 months (8 months in this embodiment) to supplement the collection of data on changes in geological defects and earthquake-prone structures. After receiving the verified detection data, the system instructs the detection mapping fusion and mapping module 4 to update the three-dimensional digital twin model and various mapping results, and to track the evolution trend of defects and changes in earthquake risk in real time. At the same time, the system interfaces the collaborative inspection results, updated mapping results, and maintenance suggestions with the subway operation and maintenance management system to output targeted operation and maintenance solutions (monitoring frequency, reinforcement suggestions, etc.) to achieve deep adaptation between the detection mapping inspection results and subway operation and maintenance. The system transmits the dynamically updated data and operation and maintenance docking records to the main control module 1 for backup.

[0101] Main control module 1 is the core of the system's control. It is electrically connected to preprocessing and area division module 2, multi-source collaborative detection module 3, detection mapping fusion and drawing module 4, collaborative inspection module 5, and dynamic update and operation and maintenance adaptation module 6. Its specific functions include: controlling the startup, operation, and collaborative work of each module; coordinating data transmission and command interaction between modules; receiving data and results (detection data, drawing results, inspection reports, etc.) output by each module to achieve integrated system control; having human-machine interaction capabilities, with a built-in touch-screen interface that can display detection progress, drawing results, and inspection data in real time, and supporting manual intervention and parameter adjustment (such as detection point spacing, update cycle, etc.); and having data storage and export capabilities, with a built-in large-capacity storage unit that can retain all detection, drawing, and inspection process data and results for easy subsequent querying, tracing, and archiving.

[0102] The specific embodiments of the present invention are not limited to the above embodiments. Those skilled in the art can flexibly adjust parameters such as the spacing between detection points, dynamic update cycle, and gravimeter model according to the actual detection scenario, subway structure parameters, and geological conditions. As long as they do not depart from the technical solution defined by the present invention, they all fall within the protection scope of the present invention.

[0103] Furthermore, the core innovation of this invention lies in multi-source collaborative detection with gravity exploration as the core, improved BP neural network data fusion, triple collaborative verification mechanism and closed-loop process. The above embodiments are only examples adapted to a specific subway operation and maintenance scenario. In the subway survey and construction phases, the application of this invention can be achieved by adjusting the detection area and the focus of detection methods, and the same or similar technical effects can be achieved.

Claims

1. A collaborative inspection method for detecting and mapping geological defects related to subway structural safety, characterized by: Includes the following steps: S1. Collect basic data or information along the subway line, delineate the core detection area and extended detection area, conduct detection environment calibration and establish an error correction model; S2. Taking gravity exploration as the core, integrating micro-motion detection, ground-penetrating radar detection, and elastic wave CT detection, collect and preprocess multi-source detection data to initially identify geological defects and earthquake-prone geological structures; S3. Integrate the preprocessed detection data through a multi-source data fusion model, draw a distribution map of layered geological defects and earthquake-prone geological structures, and perform an initial verification of the drawing accuracy. S4. Through a triple inspection mechanism, establish the correlation between geological defects, earthquake-prone geological structures and subway structural safety, evaluate the inspection results and generate feedback. S5. Regularly conduct detection and verification, update the mapping results and 3D model, connect the collaborative inspection results with the subway operation and maintenance management system, output targeted operation and maintenance suggestions, and then use the updated results as input to feed back into the error correction model in step S1 and the multi-source collaborative detection scheme in step S2, so as to realize the iterative tracking of the dynamic evolution of geological defects and the adaptive optimization of the geological model.

2. The collaborative inspection method for detecting and mapping geological defects related to subway structural safety as described in claim 1, characterized in that: In step S1, the basic data or information includes geological survey reports along the subway line, subway structural design parameters, existing geophysical exploration data, and regional seismic geological data; The core detection area is within 50m around the subway tunnel, used to investigate shallow geological defects; The extended detection area is within 100-200m outside the core detection area, used to investigate deep earthquake-prone geological structures; The method for constructing and using the error correction model is as follows: collect historical data, including basic geological parameter samples and their corresponding multi-source detection error data. Use the basic geological parameter samples as input and the corresponding detection error data as the expected output to conduct supervised training on the BP neural network, so that the network learns the mapping relationship from geological parameters to errors. The trained model can automatically predict and output the corresponding error correction parameters for new basic geological parameters.

3. The collaborative inspection method for detecting and mapping geological defects related to subway structural safety as described in claim 1, characterized in that: In step S3, the multi-source data fusion model is an improved BP neural network fusion model, the improvement being: The gravity exploration data is assigned a higher initial connection weight at the input layer, and the gravity data is normalized independently of other data sources. Draw a distribution map of layered geological defects and earthquake-prone geological structures, specifically including: generating a distribution map of shallow geological defects, a distribution map of deep earthquake-prone geological structures, and a three-dimensional comprehensive distribution map. The initial verification error of the drawing accuracy should be ≤5%.

4. The collaborative inspection method for detecting and mapping geological defects related to subway structural safety as described in claim 1, characterized in that: In step S4, the triple verification mechanism includes detection-drawing comparison verification, drawing-structural safety correlation verification, and on-site practical verification. The pass criterion for detection-drawing comparison verification is that the consistency between the two is ≥95%, and the pass criterion for drawing-structural safety correlation verification is that the prediction error of its correlation model is ≤5%. In step S5, the dynamic update cycle is every 6-12 months, and gravity exploration and micro-motion detection are carried out regularly for verification.

5. A collaborative inspection system for detecting and mapping geological defects related to subway structural safety, used to implement the method described in any one of claims 1-4, characterized in that: include: The main control module (1), and the following components electrically connected to and controlled by the main control module (1): The preprocessing and area division module (2) is used to perform step S1, including collecting basic data or information along the subway line, delineating the core detection area and the extended detection area, and establishing an error correction model based on the BP neural network. The multi-source collaborative detection module (3) is used to perform step S2, including data acquisition based on gravity exploration and integrating micro-motion detection, ground radar detection and elastic wave CT detection, as well as preprocessing the acquired multi-source detection data. The data fusion and plotting module (4) is used to perform step S3, including using an improved BP neural network fusion model to fuse the preprocessed multi-source detection data, and plotting shallow geological defect distribution map, deep earthquake-prone geological structure distribution map and three-dimensional comprehensive distribution map based on the fusion results; The collaborative inspection module (5) is used to perform step S4, including conducting detection-drawing comparison inspection, drawing-structural safety correlation inspection and on-site practical inspection, and establishing correlation relationships and forming feedback based on the inspection results; The dynamic update and operation and maintenance adaptation module (6) is used to execute step S5, including triggering detection and verification according to a preset cycle to update the drawing and model, and connecting the results to the subway operation and maintenance management system.

6. The system according to claim 5, characterized in that: The multi-source collaborative detection module (3) includes a gravity exploration unit (301), a micro-motion detection unit (302), a ground-penetrating radar detection unit (303), an elastic wave CT detection unit (304), and a data preprocessing unit (305). The gravity exploration unit (301) is configured to use a high-precision gravimeter and to set up detection points in a grid pattern to collect gravity anomaly data and gravity gradient data. The data preprocessing unit (305) is configured to perform noise reduction, correction and standardization processing on the data from each detection unit.

7. The system according to claim 6, characterized in that: The data preprocessing unit (305) performs correction operations on the data collected by the gravity exploration unit (301), specifically including terrain correction, Bouguer correction and equilibrium correction.

8. The system according to claim 5, characterized in that: The data fusion and plotting module (4) includes: The data fusion unit (401) is configured to run an improved BP neural network fusion model, which integrates multi-source data by assigning higher initial connection weights to gravity exploration data and performing independent normalization processing, and outputs the fused geological body parameters. The drawing unit (402) is configured to generate a shallow geological defect distribution map, a deep earthquake-prone geological structure distribution map, and a three-dimensional comprehensive distribution map based on the geological body parameters output by the data fusion unit (401).

9. The system according to claim 5, characterized in that: The collaborative inspection module (5) is configured as follows: The plotting results output by the data fusion and plotting module (4) are compared with the original detection data of the multi-source collaborative detection module (3) to calculate the consistency. By combining the subway structural design parameters, the associated model is called to calculate the risk probability of the geological structures marked in the drawing results to the safety of the subway structure; It receives input from on-site practical tests, comprehensively evaluates the test results, and generates feedback information including operation and maintenance suggestions.

10. The system according to claim 5, characterized in that: The dynamic update and operation and maintenance adaptation module (6) is configured as follows: Every 6 to 12 months, a command is sent to the multi-source collaborative detection module (3) to initiate gravity exploration and micro-motion detection verification; Update the drawing results and 3D model in the data fusion and drawing module (4) according to the verification data; The updated model, test results, and operation and maintenance suggestions will be pushed to the subway operation and maintenance management system. The updated drawing results are then fed back to the error correction model of the preprocessing and region division module (2) and the detection scheme configuration of the multi-source collaborative detection module (3).

Images