Intelligent detection system and method suitable for underground space composite supporting structure

Abstract

The application belongs to the field of underground construction safety detection, and provides an intelligent detection system and method suitable for underground space composite supporting structure, obtains BIM parameters and real-time monitoring data of the supporting structure, and carries out feature extraction to obtain a monitoring feature vector; BIM confidence, monitoring confidence and monitoring feature confidence are calculated respectively; data conflicts are handled according to the size relationship of the BIM confidence, the monitoring confidence and the monitoring feature confidence and a preset arbitration rule to obtain conflict-free fusion data; a supporting structure digital twin is constructed based on the conflict-free fusion data to carry out simulation, and a damage index and a surrounding rock-support contact stress distribution are obtained. The application solves the problems of data fragmentation, poor dynamic adaptability, heterogeneous data fusion conflict, weak interpretability, low generalization ability and rigid intervention in traditional underground space composite supporting detection.

Description

Technical Field

[0001] This invention belongs to the field of underground construction safety detection technology, specifically relating to an intelligent detection system and method suitable for composite support structures in underground spaces. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] As urban underground space development continues to extend deeper, composite support structures such as pipe roofs, steel grid frames, shotcrete layers, and small-diameter pipe grouting have become key safety guarantees for projects such as deep-buried tunnels and mine roadways due to their superior load-bearing capacity and adaptability to complex geological conditions. Existing technologies have attempted to integrate BIM and digital twin technologies, and introduced multinomial response surface models and attention mechanisms to assess damage during construction, forming a closed-loop detection framework, which has initially alleviated the problems of data fragmentation and reliance on experience.

[0004] However, this scheme still has the following core flaws: the authenticity of multi-source data itself is difficult to guarantee. Sensor data is susceptible to environmental interference such as high humidity and strong electromagnetic fields, and improper deployment can lead to insufficient data representativeness. The deviation between BIM design parameters and unforeseen geological conditions during construction can reach more than 20%. The fixed weights and single CNN architecture used cannot adapt to the differentiated characteristics of dynamic disturbance stages such as surrounding rock creep, blasting vibration, and sudden tunneling changes. The extracted features are easily mixed with noise or non-disturbance invalid features, resulting in low matching degree with historical damage cases and difficulty in reflecting real damage.

[0005] The fusion of heterogeneous data lacks credibility assessment and conflict arbitration mechanisms. Simply splicing or fusion with fixed weights of sensor measurements, BIM design values, and CNN features fails to quantify their individual credibility. When conflicts arise between sensor anomalies and outdated design parameters, they cannot be resolved. Direct normalization under conflicting dimensions ignores reliability differences, leading to distorted fused data. The mapping of the digital twin to the entity's state deviates from reality, causing damage assessment and decision-making to lack a reliable basis and posing a risk of misjudgment. Summary of the Invention

[0006] To address the aforementioned issues, this invention proposes an intelligent detection system and method suitable for composite support structures in underground spaces. This invention can perform credibility assessment and conflict arbitration on multi-source data, and achieve adaptive matching of feature extraction with dynamic disturbances, thereby ensuring the authenticity of the data and the accuracy of the parameters, providing a reliable foundation for the intelligent detection of composite support structures in underground spaces.

[0007] According to some embodiments, the first aspect of the present invention provides an intelligent detection method suitable for composite support structures in underground spaces, employing the following technical solution: Intelligent detection methods applicable to composite support structures in underground spaces include: Obtain the BIM parameters and real-time monitoring data of the support structure, and extract features from the real-time monitoring data to obtain the monitoring feature vector; Confidence levels are calculated for the BIM parameters, real-time monitoring data, and monitoring feature vectors of the support structure to obtain BIM confidence level, monitoring confidence level, and monitoring feature confidence level. Data conflicts are handled according to the relationship between the BIM confidence level, monitoring confidence level, and monitoring feature confidence level, as well as the preset arbitration rules, to obtain conflict-free fused data. A digital twin of the support structure is constructed based on conflict-free fused data. Simulation is performed using the digital twin of the support structure and the conflict-free fused data to obtain the damage index and the distribution of contact stress between the surrounding rock and the support. Based on the damage index and the distribution of contact stress between the surrounding rock and the support, monitoring feature anomaly markers are superimposed to generate a global stress cloud map.

[0008] Further, the step of calculating confidence levels for the BIM parameters, real-time monitoring data, and monitoring feature vectors of the support structure to obtain BIM confidence level, monitoring confidence level, and monitoring feature confidence level includes: Calculate the BIM confidence level based on the matching degree between the BIM parameters and the specification parameters of the support structure; The data integrity rate is determined based on real-time monitoring data, and the monitoring confidence level is calculated based on the data integrity rate. Calculate the cosine similarity between the monitored feature vector and the monitored feature vector in the historical damage case database, and determine the confidence level of the monitored feature based on the magnitude of the cosine similarity.

[0009] Furthermore, the step of processing data conflicts based on the relationship between BIM confidence level, monitoring confidence level, and monitoring feature confidence level, as well as preset arbitration rules, to obtain conflict-free fused data includes: The first type of conflict occurs when the value of the real-time monitoring data is greater than the BIM parameter, and the monitoring confidence level is higher than the BIM confidence level. In this case, the real-time monitoring data shall prevail, and the BIM parameter shall be marked as pending update. The second type of conflict is multi-source data dimension conflict. Dimensionless normalization is performed on each type of data source, and weighted weights are obtained through the confidence scores of the data sources. The data sources after dimensionless normalization and the weighted weights are then merged to obtain a weighted fusion weight. The third type of conflict is missing data, which uses BIM parameters as substitute parameters and reduces the basic weight of BIM parameters; Based on the different processing procedures for the three types of conflicts mentioned above, conflict-free fused data is obtained.

[0010] Furthermore, the calculation of BIM confidence based on the matching degree between the BIM parameters and the specification parameters of the support structure includes: The BIM confidence level is determined based on the magnitude and number of deviations between the BIM parameters and the specification parameters of the support structure; among which, the deviation rate... The calculation is as follows:

[0011] in, For the first Measured values ​​of each BIM parameter For the first The standard values ​​for each BIM parameter.

[0012] Furthermore, the digital twin employs a polynomial response surface model as the mechanical response prediction model, as follows;

[0013] in, For the first The basic response surface output corresponding to each output result; For conflict-free fusion data, the first A conflict-free fusion parameter, For the first The constant term corresponding to each mechanical response output; For the first The first mechanical response output The coefficients of the first-order terms of the conflict-free fusion parameters, For the first The first mechanical response output The quadratic coefficients of the conflict-free fusion parameters, For the first The first mechanical response output The first conflict-free fusion parameter and the second The cross-term response surface coefficients among the conflict-free fusion parameters; , The first The and the first One fusion input parameter; All represent the numbers of conflict-free fusion parameters, and satisfy the following conditions in the cross term: , This represents the total number of conflict-free fusion parameters.

[0014] Furthermore, the damage index The calculation formula is as follows:

[0015] in, , For the support structure in the first The actual number of load cycles borne within a range of plastic strain amplitudes; For the first The allowable number of fatigue cycles corresponding to each range of plastic strain amplitude; For the first The representative plastic strain amplitude corresponding to each range of plastic strain amplitudes, Let be the material damage constant. The damage evolution index; The range of plastic strain amplitude is numbered, and , This represents the total number of ranges of plastic strain amplitude.

[0016] According to some embodiments, the second aspect of the present invention provides an intelligent detection system suitable for composite support structures in underground spaces, employing the following technical solution: Intelligent detection systems applicable to composite support structures in underground spaces include: The multi-source data processing module is configured to acquire BIM parameters of the support structure and real-time monitoring data, and to extract features from the real-time monitoring data to obtain monitoring feature vectors. The data fusion module is configured to calculate the confidence scores of the BIM parameters, real-time monitoring data, and monitoring feature vectors of the support structure, respectively, to obtain the BIM confidence score, monitoring confidence score, and monitoring feature confidence score; and to handle data conflicts based on the relationship between the BIM confidence score, monitoring confidence score, and monitoring feature confidence score and the preset arbitration rules to obtain conflict-free fused data. The digital twin simulation module is configured to construct a digital twin of the support structure based on conflict-free fused data, and to perform simulation using the digital twin of the support structure and conflict-free fused data to obtain the damage index and the distribution of contact stress between the surrounding rock and the support. The decision optimization module is configured to generate a global stress cloud map by overlaying monitoring feature anomaly markers based on the damage index and the distribution of contact stress between the surrounding rock and the support.

[0017] According to some embodiments, a third aspect of the present invention provides a computer-readable storage medium.

[0018] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the intelligent detection method for composite support structures in underground spaces as described in the first embodiment above.

[0019] According to some embodiments, a fourth aspect of the present invention provides a computer device.

[0020] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the intelligent detection method for composite support structures in underground spaces as described in the first embodiment above.

[0021] According to some embodiments, a fifth aspect of the present invention provides a computer program product or computer program.

[0022] A computer program product or computer program includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps in the intelligent detection method for composite support structures in underground spaces as described in the first embodiment above.

[0023] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention improves dynamic adaptability and assessment accuracy, and solves the shortcomings of fixed feature extraction. Through an adaptive feature selection mechanism for disturbance stages, it identifies disturbance stages such as the rock creep stage, blasting vibration stage, and tunneling abrupt change stage, dynamically switching between 1DCNN and 2DCNN architectures and adjusting attention weights based on CNN feature vector values. This mechanism breaks through the limitations of traditional fixed weights and a single architecture, accurately capturing local anomalies at different stages (such as spatial strain concentration during creep and temporal vibration peaks during blasting), upgrading damage assessment from "macro-average" to "micro-precise focusing," significantly improving assessment accuracy in complex construction scenarios.

[0024] This invention achieves interpretable synchronization and transparent traceability, constructing a precise local-global association. Through a bidirectional synchronization mapping mechanism, it extends unique identifiers (UUIDs) to CNN features and dynamic attention weights, establishing a full-link association mapping table for BIM nodes, sensors, features, and weights. A knowledge graph is used to store the relationships between features, weights, damage, and arbitration results. This mechanism overcomes the black-box problem of traditional synchronization, allowing engineers to intuitively understand how local anomalies affect the global structure and build trust in the model.

[0025] This invention enhances the precision and generalization of intervention, shifting risk prevention and control from passive to proactive. Through trend-driven prevention and control and a two-dimensional calibration mechanism, it replaces fixed threshold-based tiered responses with damage growth rate and acceleration, combined with a geological-construction two-dimensional model calibration to adapt to different geological conditions and construction methods. This mechanism avoids over-intervention or under-intervention, achieving an upgrade from post-event emergency response to pre-event prediction, significantly improving the proactiveness and universality of risk prevention and control. Attached Figure Description

[0026] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0027] Figure 1 This is a flowchart of an intelligent detection method applicable to composite support structures in underground spaces, as described in this embodiment of the invention. Detailed Implementation

[0028] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0029] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0030] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0031] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0032] Example 1 like Figure 1 As shown, this embodiment provides an intelligent detection method applicable to composite support structures in underground spaces. This embodiment uses the application of this method to a server as an example for illustration. It is understood that this method can also be applied to terminals, and can also be applied to systems including terminals, servers, and other components, and is implemented through interaction between the terminal and the server. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network servers, cloud communication, middleware services, domain name services, CDN security services, and big data and artificial intelligence platforms. The terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, etc., but is not limited to these. The terminal and server can be directly or indirectly connected via wired or wireless communication, which is not limited herein. In this embodiment, the method includes the following steps: Step S1: Obtain the BIM parameters of the support structure and real-time monitoring data, and extract features from the real-time monitoring data to obtain the monitoring feature vector; Step S1.1: Parse the structured parameter file through the BIM interface to obtain the BIM parameters of the support structure; Specifically, the first step is to extract and store the BIM parameters of the support structure. By parsing the structured parameter file through the BIM interface, the geometric parameters, material constitutive parameters, design loads and construction disturbance parameters of the support structure are extracted and stored as structured data files to build a macroscopic skeleton for the digital twin.

[0033] Step S1.2: Use sensors installed at various parts of the support structure to obtain real-time monitoring data of the support structure.

[0034] Conduct real-time monitoring data acquisition and sensor deployment. Install sensors at key support locations, and deploy rebar gauges at nodes and welds of pipe sheds and grating steel frames to monitor stress changes in the steel frame; The surrounding rock pressure gauge is embedded at the contact surface between the shotcrete layer and the surrounding rock, and around the grouting area of ​​the small pipe, to capture the distribution of surrounding rock pressure. The convergence gauge is installed at the top of the tunnel and the middle of the sidewalls to measure the clearance deformation. Strain gauges are embedded inside the shotcrete layer to monitor the accumulation of micro-strain; piezometers are buried near small conduits in the aquifer to collect pore water pressure.

[0035] The sensor is connected to the field data acquisition module via cable, and the monitoring data is uploaded to the cloud server in real time via the MQTT protocol.

[0036] Step S1.3: Select the corresponding feature extraction method based on different sensors, extract features from the real-time monitoring data, and obtain the monitoring feature vector.

[0037] 1D CNN is used to process time series data such as convergence displacement curves and steel bar stress time histories to extract local anomaly features such as displacement peaks and stress surge areas. 2D CNN is used to process spatial data such as the distribution matrix of surrounding rock pressure gauges and the spatial layout of strain gauges to identify local concentrated areas such as strain concentration areas and blasting peak areas.

[0038] Based on the corresponding feature extraction method, features are extracted from the real-time monitoring data to obtain the monitoring feature vector.

[0039] Step S2: Calculate the confidence scores for the BIM parameters, real-time monitoring data, and monitoring feature vectors of the support structure to obtain the BIM confidence score, monitoring confidence score, and monitoring feature confidence score; handle data conflicts according to the relationship between the BIM confidence score, monitoring confidence score, and monitoring feature confidence score and the preset arbitration rules to obtain conflict-free fused data; Step S2.1: Calculate the BIM confidence level based on the matching degree between the BIM parameters and the specification parameters of the support structure; BIM confidence level The calculation is based on the matching degree between the BIM parameters and the code parameters of the support structure; the scoring is based on the code matching degree, and the rules are as follows: The optimal BIM parameter matching score for full compliance with the specification parameters is 0.9. When BIM parameters match the standard parameters, there is one item with an error rate of less than or equal to 10%, which is 0.7. When BIM parameters match the standard parameters, the value is 0.3 if the number of out-of-tolerance items is greater than 1 or the out-of-tolerance rate is greater than 10%. Among them, 0.9 / 0.7 / 0.3 is the preferred score based on the engineering risk level: low-risk projects can be adjusted to 0.95 / 0.8 / 0.5, and high-risk projects can be tightened to 0.85 / 0.6 / 0.2.

[0040] Specifically, deviation refers to the comparison between BIM parameters and the allowable deviations in the technical specifications, and its deviation rate is... Defined as:

[0041] in, For the first Measured values ​​of each BIM parameter For the first The standard values ​​for each BIM parameter.

[0042] Step S2.2: Determine the data integrity rate based on real-time monitoring data, and calculate the monitoring confidence level based on the data integrity rate. The formula is as follows:

[0043] in, These are weighting coefficients, preferably 0.6 and 0.4; This refers to data integrity rate, for example, the data integrity rate over the past hour; It is a time-sensitive factor; Data integrity rate is obtained based on real-time monitoring data, within a preset time period. The theoretical sampling quantity for real-time monitoring data is: The first one actually received and retained after outlier removal The effective sampling quantity of real-time monitoring data is The overall data completeness rate is:

[0044] in, It represents the total number of types of real-time monitoring data; Step S2.3: Calculate the cosine similarity between the monitored feature vector and the monitored feature vector in the historical damage case database, and determine the confidence level of the monitored feature based on the magnitude of the cosine similarity; The CNN features are scored using cosine similarity comparison against a historical damage case database. This database stores at least the monitoring data, disturbance stage labels, damage state labels, and corresponding feature vectors from past projects, which are then compared with the current monitoring feature vectors.

[0045] Step S2.4: Based on the relationship between BIM confidence level, monitoring confidence level, and monitoring feature confidence level, as well as the preset arbitration rules, data conflicts are processed to obtain conflict-free fused data; Conflicts are resolved through the arbitration rules table: Conflict 1 occurs when the sensor value exceeds the BIM design value and the monitoring confidence level is higher than the BIM confidence level. In this case, the real-time monitoring data shall prevail, the BIM parameters shall be marked as pending updates, and an API alarm shall be issued. Conflict 2 is a conflict of dimensions between multiple data sources. For example, when pressure (MPa) and displacement (mm) need to be merged, firstly, dimensionless normalization is performed. The formula for dimensionless normalization is:

[0046] in, for These correspond to three types of data sources: BIM parameters, real-time monitoring data, and monitoring feature vectors. For the first The raw data of the data source. and These are the minimum and maximum values ​​in the design specifications for this type of data, respectively.

[0047] Then, the corresponding weighted weights of BIM parameters are obtained by using confidence scores from various data sources. Real-time monitoring data weighting , monitoring feature vector weighting The calculation is as follows:

[0048]

[0049]

[0050] Among them, the basic weights of BIM parameters Real-time monitoring data basic weight Monitoring the basic weights of feature vectors The calculation is as follows:

[0051] in, , Corresponding to three types of data sources—BIM parameters, real-time monitoring data, and monitoring feature vectors—the system pre-assigns a basic level of importance to each data source. , , . , , These are BIM confidence level, monitoring confidence level, and monitoring feature confidence level, respectively. , , The three data sources are assigned base weights to reflect the pre-defined importance of different data sources in engineering inspection. Then, the weighted fused data is obtained through a formula. The calculation is as follows:

[0052] Finally, the weighted fused data is input into the model; When conflict 3 is due to missing data, BIM parameters are used as substitute parameters, and the basic weight of BIM parameters is reduced by 10% to 30%, preferably by 20%, and the missing data is marked. If a conflict in the dimensions of multi-source data occurs again, the weighted weights will be recalculated based on the reduced BIM base weights, and conflict-free fused data will be output.

[0053] The reason for reducing the base weights of BIM parameters is that when the measured data is normal, BIM parameters can be continuously verified by on-site monitoring; once the measured data is missing, the system loses its ability to directly observe the true state of the current structure. If the original weights of BIM are maintained at this time, it is equivalent to assuming the design state is approximately equal to the current state, but the actual state does not match the default.

[0054] Step S3: Construct a digital twin of the support structure based on conflict-free fused data, and use the digital twin of the support structure and conflict-free fused data to perform simulation to obtain the damage index and the distribution of contact stress between the surrounding rock and the support. The conflict-free fusion data includes BIM parameters, real-time monitoring data, and monitoring feature vectors. BIM parameters are used to determine the spatial location, component type, material parameters, design load, and rock-support contact area of ​​the support structure. Real-time monitoring data reflects the current actual state of the support structure, including stress, displacement, strain, rock pressure, and pore water pressure. The monitoring feature vectors are local anomaly features extracted from the real-time monitoring data by the CNN in step S1, used to characterize abnormal changes in local areas of the support structure.

[0055] The digital twin first establishes a spatial model of the support structure based on BIM parameters, determining the positional relationships between the pipe roof, the grid steel frame, the shotcrete layer, the grouting area of ​​the small pipes, and the rock-support contact area. Then, real-time monitoring data and monitoring feature vectors are mapped to the corresponding support components or rock-support contact areas to correct the load state, deformation state, stress state, and material state of each area.

[0056] In the digital twin, a polynomial response surface model is used as the mechanical response prediction model. This polynomial response surface model is used to establish the mapping relationship between conflict-free fused data and the mechanical response of the support structure.

[0057] The input to the polynomial response surface model is conflict-free fused data:

[0058] in, For conflict-free data fusion; For the first The conflict-free fusion parameters include BIM parameters, real-time monitoring parameters, and normalized monitoring feature components.

[0059] For the The mechanical response output, represented by a polynomial response surface model, can be expressed as:

[0060] in, For the first The basic response surface output corresponding to each output result; For the first The constant term corresponding to each mechanical response output; For the first The first mechanical response output The coefficients of the first-order terms of the conflict-free fusion parameters, For the first The first mechanical response output The quadratic coefficients of the conflict-free fusion parameters, For the first The first mechanical response output The first conflict-free fusion parameter and the second The cross-term response surface coefficients among the conflict-free fusion parameters; , The first The and the first One fusion input parameter; All represent the numbers of conflict-free fusion parameters, and satisfy the following conditions in the cross term: , This represents the total number of conflict-free fusion parameters.

[0061] No. Each mechanical response output includes the plastic strain amplitude of each support zone, the actual number of load cycles, and the contact stress value between the surrounding rock and the support.

[0062] Among them, the plastic strain amplitude and the actual load cycle number are used to calculate the damage index of the support structure; the rock-support contact stress values ​​are summarized and mapped according to their corresponding BIM spatial locations to form the rock-support contact stress distribution.

[0063] After obtaining the plastic strain amplitude and actual load cycle number of each support area from the digital twin, the data is substituted into the damage index calculation formula to obtain the damage index of the support structure.

[0064] Damage Index The calculation formula is as follows:

[0065] in, , For the support structure in the first The actual number of load cycles borne within a range of plastic strain amplitudes; For the first The allowable number of fatigue cycles corresponding to each range of plastic strain amplitude; For the first The representative plastic strain amplitude corresponding to a plastic strain amplitude range, wherein the representative plastic strain amplitude can be taken as the median, mean or equivalent plastic strain amplitude of the plastic strain amplitude range; Let be the material damage constant. The damage evolution index; preferred The value ranges from 0.1 to 0.4. The value ranges from 0.3 to 0.6. The range of plastic strain amplitude is numbered, and , This represents the total number of ranges of plastic strain amplitude.

[0066] At the same time, the digital twin summarizes and maps the contact stress values ​​of each rock-support contact area according to the BIM spatial location, forming the rock-support contact stress distribution.

[0067] The stress distribution in the surrounding rock-support contact area can be expressed as:

[0068] in, This indicates the stress distribution in the surrounding rock-support contact area; Indicates the first Contact stress values ​​in the surrounding rock-support contact area; This indicates the total number of contact areas between the surrounding rock and the support. Therefore, the output of step S3 includes two parts: One is the damage index, which is used to characterize the overall cumulative damage level of the support structure; Second, it is used to characterize the stress distribution of the surrounding rock-support contact area in different rock-support contact zones.

[0069] Step S4: Establish a data mapping table for the universally unique identifiers corresponding to BIM parameters, real-time monitoring data, monitoring feature vectors, and attention weight assignments; A data mapping table is established using unique identifiers (UUIDs) for BIM nodes, sensors, CNN features such as Local-Feature-001 (strain concentration area), and dynamic attention weights such as Attention-Weight-001 (surrounding rock creep assignment ID). This table includes confidence scores, conflict arbitration results, and weight adjustment records. This enables a closed loop from API sensors to CNN features to twin calibration to BIM updates. A knowledge graph stores the relationships between features, weights, damage, and arbitration results, supporting click-based traceability within a WebGL interface. The data mapping table is built using raw sensor sampling data as the underlying foundation nodes.

[0070] Specifically, each raw monitoring data node is assigned a corresponding universally unique identifier (UUID). The raw monitoring data includes at least: sensor type; sensor installation location; sampling time; actual monitoring value; and data confidence level.

[0071] Among them, a spatial mapping relationship is established between the original monitoring data nodes and the corresponding BIM component nodes to determine the location of the support structure corresponding to the monitoring data; After the original monitoring data nodes are processed by CNN feature extraction, corresponding monitoring feature nodes are generated; the monitoring feature nodes include: feature type; feature vector; abnormal region marker; feature confidence level.

[0072] Based on the monitoring feature nodes, corresponding dynamic attention weight nodes are generated. These dynamic attention weight nodes are used to characterize the importance of each local feature under the current perturbation stage.

[0073] The monitoring feature nodes and dynamic attention weight nodes are input into the digital twin model to generate damage result nodes and surrounding rock-support contact stress nodes, and then associated with the corresponding BIM component nodes.

[0074] This forms a full-link mapping relationship between the original monitoring data node, the local feature node, the dynamic attention weight node, the digital twin result node, and the BIM component node.

[0075] The data mapping table also records: data confidence level; conflict arbitration results; weight adjustment records; BIM update records; and damage assessment results.

[0076] Furthermore, a knowledge graph is used to store the relationships between the aforementioned nodes to support click tracing and fully interpretable analysis of the WebGL interface.

[0077] This step enables virtual-real linkage by assigning UUIDs to BIM nodes, sensors, CNN features, and dynamic attention weights, and establishing a mapping table containing confidence scores, conflict arbitration results, and weight adjustment records. The API is used to achieve a closed loop of sensor → CNN features → twin calibration → BIM update.

[0078] Step S5: Based on the damage index and the distribution of contact stress between the surrounding rock and the support, superimpose monitoring feature anomaly markers to generate a global stress cloud map. Based on the global stress cloud map, predict the remaining life of the support structure through a linear degradation model, and generate maintenance recommendations based on the predicted remaining life.

[0079] The global stress cloud map is rendered using WebGL. The von Mises stress is mapped using a rainbow spectrum and the microcracks are visualized using a thermal map with a damage hotspot map. Local anomaly markers such as red boxes are superimposed to indicate local strain concentration areas. The remaining life is predicted by a linear degradation model, a trend-driven hierarchical prevention and control strategy is adopted, and maintenance suggestions are generated by calibrating a geological-construction dual-dimensional model. The BIM maintenance plan is automatically updated and fed back to the parameter extraction step to form a closed-loop iteration.

[0080] In other words, a trend-driven, tiered prevention and control strategy is adopted, using damage growth rate ΔD / Δt and acceleration Δ(ΔD / Δt) / Δt to classify low, medium, and high risk responses. The preferred low-risk response is to maintain parameters when the growth rate is less than 0.05 per day and the acceleration is less than 0.01 per day squared. The medium-risk response is to reduce the speed and increase the frequency of monitoring when the growth rate is less than 0.1 per day or the acceleration is greater than or equal to 0.01 per day squared. The high-risk response is to suspend reinforcement when the growth rate is greater than or equal to 0.1 per day or D is greater than or equal to 0.8 and the growth rate is greater than or equal to 0.05 per day. The system combines a geological-construction dual-dimensional model to calibrate preset scenario-parameter tables for soft rock + drill-and-blast method and hard rock + shield tunneling method, generates maintenance suggestions, automatically updates the BIM maintenance plan, and feeds back to the parameter extraction step to form a closed loop.

[0081] The entire implementation process aims at risk prevention and control during the construction period. By accurately deploying sensors to capture data from key parts, and through cloud-based intelligent processing, digital twin assessment, two-way synchronous mapping, and visualized decision-making, intelligent management of the entire process from disturbance monitoring to proactive prevention and control is achieved, ensuring the safety of the support structure.

[0082] This example addresses the adaptability issue of fixed feature extraction by using adaptive feature selection during the perturbation phase to identify the dynamic switching of CNN architecture and attention weights during perturbation phases. Confidence arbitration fusion achieves reliable fusion of heterogeneous data through multi-dimensional confidence scoring and conflict arbitration rules, avoiding data distortion. Bidirectional synchronous mapping extends UUIDs to CNN features and dynamic weights, establishing a full-link association between nodes, sensors, features, and weights, and enabling interpretable traceability using knowledge graphs and APIs. Trend-driven prevention and control, along with two-dimensional calibration, replaces fixed threshold-based graded responses with damage evolution trends, combining geological and construction scenarios to calibrate model parameters, improving intervention accuracy and generalization capabilities.

[0083] This embodiment aims to eliminate data fragmentation to achieve closed-loop management, improve dynamic adaptability and evaluation accuracy, enhance the credibility and interpretability of data fusion, strengthen generalization ability and intervention precision, promote the upgrade of detection mode from passive emergency response to intelligent prediction, and significantly improve the safety and intelligence level of support structure during the construction period of deep underground space.

[0084] Example 2 This embodiment provides an intelligent detection system suitable for composite support structures in underground spaces, including: The multi-source data processing module is configured to acquire BIM parameters of the support structure and real-time monitoring data, and to extract features from the real-time monitoring data to obtain monitoring feature vectors. The data fusion module is configured to calculate the confidence scores of the BIM parameters, real-time monitoring data, and monitoring feature vectors of the support structure, respectively, to obtain the BIM confidence score, monitoring confidence score, and monitoring feature confidence score; and to handle data conflicts based on the relationship between the BIM confidence score, monitoring confidence score, and monitoring feature confidence score and the preset arbitration rules to obtain conflict-free fused data. The digital twin simulation module is configured to construct a digital twin of the support structure based on conflict-free fused data, and to perform simulation using the digital twin of the support structure and conflict-free fused data to obtain the damage index and the distribution of contact stress between the surrounding rock and the support. The decision optimization module is configured to generate a global stress cloud map by overlaying monitoring feature anomaly markers based on the damage index and the distribution of contact stress between the surrounding rock and the support.

[0085] The examples and application scenarios implemented by the above modules and corresponding steps are the same, but are not limited to the content disclosed in Embodiment 1 above. It should be noted that the above modules, as part of the system, can be executed in a computer system such as a set of computer-executable instructions.

[0086] The descriptions of each embodiment in the above embodiments have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0087] The proposed system can be implemented in other ways. For example, the system embodiments described above are merely illustrative, and the division of modules described above is only a logical functional division. In actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed.

[0088] Example 3 This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the intelligent detection method for composite support structures in underground spaces as described in Embodiment 1 above.

[0089] Example 4 This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the intelligent detection method for composite support structures in underground spaces as described in Embodiment 1 above.

[0090] Example 5 This embodiment provides a computer program product or computer program, including computer instructions stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps in the intelligent detection method for underground space composite support structures described in Embodiment 1 above.

[0091] Those skilled in the art will understand that embodiments of the present invention can provide methods, systems, or computer program products. Therefore, the present invention can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0092] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart. Figure 1 One or more processes and / or boxes Figure 1A device that provides the functions specified in one or more boxes.

[0093] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0094] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0095] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0096] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. An intelligent detection method applicable to composite support structures in underground spaces, characterized in that, include: Obtain the BIM parameters and real-time monitoring data of the support structure, and extract features from the real-time monitoring data to obtain the monitoring feature vector; Confidence levels were calculated for the BIM parameters, real-time monitoring data, and monitoring feature vectors of the support structure, respectively, to obtain BIM confidence levels, monitoring confidence levels, and monitoring feature confidence levels, including: Calculate the BIM confidence level based on the matching degree between the BIM parameters and the specification parameters of the support structure; The data integrity rate is determined based on real-time monitoring data, and the monitoring confidence level is calculated based on the data integrity rate. Calculate the cosine similarity between the monitored feature vector and the monitored feature vector in the historical damage case database, and determine the confidence level of the monitored feature based on the magnitude of the cosine similarity; Based on the relationship between BIM confidence level, monitoring confidence level, and monitoring feature confidence level, and according to preset arbitration rules, data conflicts are handled to obtain conflict-free fused data, including: The first type of conflict occurs when the value of the real-time monitoring data is greater than the BIM parameter, and the monitoring confidence level is higher than the BIM confidence level. In this case, the real-time monitoring data shall prevail, and the BIM parameter shall be marked as pending update. The second type of conflict is multi-source data dimension conflict. Dimensionless normalization is performed on each type of data source, and weighted weights are obtained through the confidence scores of the data sources. The data sources after dimensionless normalization and the weighted weights are then merged to obtain a weighted fusion weight. The third type of conflict is missing data, which uses BIM parameters as substitute parameters and reduces the basic weight of BIM parameters; Based on the different processing procedures for the three conflict types mentioned above, conflict-free fused data is obtained; A digital twin of the support structure is constructed based on conflict-free fused data. Simulation is performed using the digital twin of the support structure and the conflict-free fused data to obtain the damage index and the distribution of contact stress between the surrounding rock and the support. Based on the damage index and the distribution of contact stress between the surrounding rock and the support, monitoring feature anomaly markers are superimposed to generate a global stress cloud map.

2. The intelligent detection method for composite support structures in underground spaces as described in claim 1, characterized in that, The calculation of BIM confidence level based on the matching degree between BIM parameters and specification parameters of the support structure includes: The BIM confidence level is determined based on the magnitude and number of deviations between the BIM parameters and the specification parameters of the support structure; among which, the deviation rate... The calculation is as follows: in, For the first Measured values ​​of each BIM parameter For the first The standard values ​​for each BIM parameter.

3. The intelligent detection method for composite support structures in underground spaces as described in claim 1, characterized in that, The digital twin uses a polynomial response surface model as the mechanical response prediction model, as follows; in, For the first The basic response surface output corresponding to each output result; For conflict-free fusion data, the first A conflict-free fusion parameter, For the first The constant term corresponding to each mechanical response output; For the first The first mechanical response output The coefficients of the first-order terms of the conflict-free fusion parameters, For the first The first mechanical response output The quadratic coefficients of the conflict-free fusion parameters, For the first The first mechanical response output The first conflict-free fusion parameter and the second The cross-term response surface coefficients among the conflict-free fusion parameters; , The first The and the first One fusion input parameter; All represent the numbers of conflict-free fusion parameters, and satisfy the following conditions in the cross term: , This represents the total number of conflict-free fusion parameters.

4. The intelligent detection method for composite support structures in underground spaces as described in claim 1, characterized in that, Damage Index The calculation formula is as follows: in, , For the support structure in the first The actual number of load cycles borne within a range of plastic strain amplitudes; For the first The allowable number of fatigue cycles corresponding to each range of plastic strain amplitude; For the first The representative plastic strain amplitude corresponding to each range of plastic strain amplitudes, Let be the material damage constant. The damage evolution index; The range of plastic strain amplitude is numbered, and , This represents the total number of ranges of plastic strain amplitude.

5. An intelligent detection system applicable to composite support structures in underground spaces, comprising performing the steps of the intelligent detection method for composite support structures in underground spaces as described in any one of claims 1-4, characterized in that, include: The multi-source data processing module is configured to acquire BIM parameters of the support structure and real-time monitoring data, and to extract features from the real-time monitoring data to obtain monitoring feature vectors. The data fusion module is configured to calculate the confidence scores of the BIM parameters, real-time monitoring data, and monitoring feature vectors of the support structure, respectively, to obtain the BIM confidence score, monitoring confidence score, and monitoring feature confidence score; and to handle data conflicts based on the relationship between the BIM confidence score, monitoring confidence score, and monitoring feature confidence score and the preset arbitration rules to obtain conflict-free fused data. The digital twin simulation module is configured to construct a digital twin of the support structure based on conflict-free fused data, and to perform simulation using the digital twin of the support structure and conflict-free fused data to obtain the damage index and the distribution of contact stress between the surrounding rock and the support. The decision optimization module is configured to generate a global stress cloud map by overlaying monitoring feature anomaly markers based on the damage index and the distribution of contact stress between the surrounding rock and the support.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the intelligent detection method for composite support structures in underground spaces as described in any one of claims 1-4.

7. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the intelligent detection method for composite support structures in underground spaces as described in any one of claims 1-4.

8. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the steps in the intelligent detection method for composite support structures in underground spaces as described in any one of claims 1-4.

Images