Tunnel surrounding rock deformation monitoring method and system

Abstract

This invention discloses a method and system for monitoring tunnel surrounding rock deformation, belonging to the field of surrounding rock deformation monitoring technology. The method includes: interactively obtaining a set of surrounding rock state associations; configuring monitoring equipment to generate a surrounding rock deformation monitoring array; pre-constructing a surrounding rock deformation monitoring cloud platform; performing source analysis based on the surrounding rock deformation early warning to obtain predicted defects in construction technology; locating construction deviations to obtain predicted construction deviation results; generating construction early warning commands and sending the commands to the construction operation and maintenance center. This invention solves the technical problems of existing technologies, such as the difficulty in real-time and comprehensive automated monitoring of surrounding rock deformation during tunnel construction, low early warning accuracy, lack of effective guarantee for tunnel construction safety, and resulting in low construction efficiency. It achieves automated monitoring and accurate early warning of surrounding rock deformation during tunnel construction, improving the safety and overall construction efficiency of tunnel construction under complex geological conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of surrounding rock deformation monitoring, and specifically relates to a method and system for monitoring the deformation of tunnel surrounding rock. Background Art

[0002] In the field of tunnel engineering construction, as construction projects continuously expand into complex geological areas, many limitations of traditional surrounding rock monitoring technologies have become increasingly apparent. Due to the weak cementation, uncementation, and extremely soft rock quality of the formation, and the complex water stability characteristics, the surrounding rock often undergoes rapid plastic deformation and rheology during construction, with extremely poor stability and is extremely prone to water and sand inrushes. However, traditional monitoring technologies are difficult to detect the subtle changes in the surrounding rock in a timely and accurate manner, and cannot provide effective early warnings for construction decisions. In terms of construction technology, there is a lack of systematic means for analyzing the seepage field, making it difficult to accurately grasp the influence of groundwater dynamics, the accuracy of advanced geological forecasting is insufficient, and the timing and method of advanced support are relatively blind, often leading to frequent safety accidents. Moreover, in the construction safety guarantee link, the operation specifications are not clear, such as the key points for hidden danger investigation and cleaning during support, lining, and tunnel body excavation are not clear, greatly affecting construction safety. At the same time, due to the lack of in-depth research on the physical and mechanical properties of the formation, the construction efficiency is quite low, and the advantages of equipment such as TBM are difficult to be fully utilized.

[0003] The prior art has technical problems such as being difficult to perform real-time and comprehensive automatic monitoring on the deformation of surrounding rock during tunnel construction, having relatively low early warning accuracy, lacking effective guarantee for the safety of tunnel construction, and resulting in low construction efficiency. Summary of the Invention

[0004] This application provides a method and system for monitoring the deformation of tunnel surrounding rock, which is used to solve the technical problems existing in the prior art, such as being difficult to perform real-time and comprehensive automatic monitoring on the deformation of surrounding rock during tunnel construction, having relatively low early warning accuracy, lacking effective guarantee for the safety of tunnel construction, and resulting in low construction efficiency.

[0005] In view of the above problems, this application provides a method and system for monitoring the deformation of tunnel surrounding rock.

[0006] In the first aspect of this application, a method for monitoring the deformation of tunnel surrounding rock is provided, and the method includes:

[0007] An interactive method is used to obtain a set of surrounding rock state correlations, which includes multiple surrounding rock state correlation indicators. Based on the set of surrounding rock state correlations, monitoring equipment is configured in the tunnel construction area to generate a surrounding rock deformation monitoring array, which includes multiple surrounding rock state monitoring devices corresponding to the multiple surrounding rock state correlation indicators. A surrounding rock deformation monitoring cloud platform is pre-constructed, and a communication link is established between the surrounding rock deformation monitoring array and the surrounding rock deformation monitoring cloud platform. When the surrounding rock deformation monitoring cloud platform receives and analyzes the feedback data from the surrounding rock deformation monitoring array and outputs a surrounding rock deformation warning, a source analysis is performed based on the surrounding rock deformation warning to obtain predicted defects in construction technology. Based on the predicted defects in construction technology, construction deviations are located to obtain predicted construction deviation results. A construction warning command is generated based on the predicted construction deviation results and sent to the construction operation and maintenance center.

[0008] A second aspect of this application provides a tunnel surrounding rock deformation monitoring system, the system comprising:

[0009] The system includes the following modules: a rock condition association set acquisition module for interactively obtaining a rock condition association set, which includes multiple rock condition association indicators; a rock deformation monitoring array generation module for configuring monitoring equipment in the tunnel construction area according to the rock condition association set, generating a rock deformation monitoring array, which includes multiple rock condition monitoring devices corresponding to the multiple rock condition association indicators; a communication link construction module for pre-constructing a rock deformation monitoring cloud and constructing a communication link between the rock deformation monitoring array and the rock deformation monitoring cloud; a construction technology prediction defect acquisition module for performing source analysis based on the rock deformation warning and obtaining construction technology prediction defects when the rock deformation monitoring cloud receives and analyzes the feedback data from the rock deformation monitoring array and outputs a rock deformation warning; a construction deviation prediction result acquisition module for locating construction deviations based on the construction technology prediction defects and obtaining construction deviation prediction results; and a construction warning instruction generation module for generating construction warning instructions based on the construction deviation prediction results and sending the construction warning instructions to the construction operation and maintenance center.

[0010] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0011] An interactive method is used to obtain a surrounding rock state correlation set, which includes multiple surrounding rock state correlation indicators. Based on this set, monitoring equipment is configured in the tunnel construction area to generate a surrounding rock deformation monitoring array, which includes multiple surrounding rock state monitoring devices corresponding to the multiple surrounding rock state correlation indicators. A surrounding rock deformation monitoring cloud platform is pre-constructed, and a communication link is established between the monitoring array and the cloud platform. When the cloud platform receives and analyzes the feedback data from the monitoring array and outputs a surrounding rock deformation warning, a source analysis is performed based on the warning to obtain predicted defects in construction technology. Construction deviations are located based on these defects to obtain predicted deviation results. A construction warning command is generated based on the predicted deviation results and sent to the construction operation and maintenance center. This method achieves automated monitoring and accurate warning of surrounding rock deformation during tunnel construction, improving the safety and overall construction efficiency of tunnel construction under complex geological conditions. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a schematic flowchart of a tunnel surrounding rock deformation monitoring method provided in an embodiment of this application;

[0014] Figure 2 This is a schematic diagram of a tunnel surrounding rock deformation monitoring system provided in an embodiment of this application.

[0015] Explanation of reference numerals in the attached figures: 10 for obtaining the surrounding rock state association set, 20 for generating the surrounding rock deformation monitoring array, 30 for constructing the communication link, 40 for obtaining the construction technology prediction defect, 50 for obtaining the construction deviation prediction result, and 60 for generating the construction early warning instruction. Detailed Implementation

[0016] This application provides a method and system for monitoring the deformation of surrounding rock in tunnels, which addresses the technical problems of existing technologies, such as the difficulty in real-time and comprehensive automated monitoring of surrounding rock deformation during tunnel construction, low accuracy of early warning, lack of effective guarantee of tunnel construction safety, and low construction efficiency.

[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0018] Example 1, as Figure 1 As shown, this application provides a method for monitoring the deformation of surrounding rock in tunnels, the method comprising:

[0019] Step S100: Interact to obtain the surrounding rock state association set, wherein the surrounding rock state association set includes multiple surrounding rock state association indicators.

[0020] Specifically, through multi-channel interaction, a set of key information that comprehensively and accurately reflects the state of the surrounding rock is obtained, namely, the surrounding rock state correlation set. This set encompasses multiple surrounding rock state correlation indicators, which reveal the characteristics and state of the surrounding rock from different dimensions. For example, the displacement index of the surrounding rock, through professional displacement monitoring equipment, can accurately measure the distance and trend of the surrounding rock movement in different directions. Whether it is a tiny millimeter-level displacement or a centimeter-level or even larger movement that may occur under complex geological stress, it can be keenly captured, thus intuitively presenting the changes in the stability of the surrounding rock. The strain index utilizes strain sensors to meticulously monitor the degree of deformation of the internal structure of the surrounding rock under external pressure, reflecting the stress response of the surrounding rock at the microscopic level, and helping to determine whether the surrounding rock is in the elastic deformation range or has approached the plastic deformation stage. The crack propagation index, with the help of crack monitoring instruments and other equipment, closely monitors the generation, extension direction, and propagation speed of cracks on the surface and inside the surrounding rock, which is of key significance for assessing the structural integrity of the surrounding rock and predicting potential collapse risks. In addition, it may include the moisture content of the surrounding rock, as changes in moisture can affect the mechanical properties of the surrounding rock in complex and weak strata, thus affecting its stability; and the sound wave propagation velocity of the surrounding rock, which indirectly infers the density and internal structural changes of the surrounding rock through the propagation characteristics of sound waves in the surrounding rock. By interactively integrating these multifaceted surrounding rock condition correlation indicators, a complete set of surrounding rock condition correlations is formed, providing indispensable basic data support for subsequent accurate monitoring, scientific analysis, and effective decision-making, ensuring that tunnel construction can proceed steadily in a safe and controllable manner.

[0021] Step S200: Configure monitoring equipment in the tunnel construction area according to the surrounding rock state association set to generate a surrounding rock deformation monitoring array, wherein the surrounding rock deformation monitoring array includes multiple surrounding rock state monitoring devices corresponding to the multiple surrounding rock state association indicators.

[0022] Specifically, based on the carefully acquired rock condition correlation set obtained in the early stages, a systematic and professional monitoring equipment configuration will be carried out in the tunnel construction area. For each rock condition correlation index in the correlation set, a highly compatible monitoring device will be precisely deployed, thus constructing a powerful rock deformation monitoring array. Taking displacement index as an example, high-precision laser displacement sensors or total stations will be installed at key locations in the tunnel surrounding rock. These devices, with their excellent measurement accuracy and stability, can capture real-time displacement changes of the surrounding rock in all directions. For strain index, strain gauges or fiber optic strain sensors will be embedded inside the surrounding rock to accurately sense the degree of deformation of the microstructure of the surrounding rock during stress, converting the stress and strain information inside the surrounding rock into electrical or optical signals for transmission. For crack propagation index, crack gauges or high-definition camera monitoring equipment will be placed in areas where cracks may occur. Crack gauges can accurately measure the dynamic changes in crack width, while camera equipment can record the development trajectory of cracks in image form, providing intuitive evidence for analyzing the causes and propagation trends of cracks. Furthermore, if the correlation dataset includes humidity indicators, humidity sensors will be distributed across different locations within the tunnel to monitor changes in the humidity of the surrounding rock in real time, providing data support for assessing the impact of humidity on rock stability. For the sound wave propagation speed indicator, sound wave transmitting and receiving devices will be strategically positioned. By emitting sound waves and measuring their propagation time and wave speed changes within the surrounding rock, the density, integrity, and internal structure changes of the rock can be indirectly determined. In this way, each rock condition monitoring device performs its specific function, ensuring that any changes in the state of the surrounding rock can be monitored promptly and accurately, effectively guaranteeing tunnel construction safety.

[0023] Step S300: Pre-build the surrounding rock deformation monitoring cloud and establish the communication link between the surrounding rock deformation monitoring array and the surrounding rock deformation monitoring cloud.

[0024] Specifically, a pre-constructed cloud platform for monitoring surrounding rock deformation is built upon advanced cloud computing technology and a big data processing architecture, possessing powerful data storage, analysis, and processing capabilities. To construct this cloud platform, a deep analysis of the surrounding rock state correlation set is required. Through index linkage effect analysis, the hidden intrinsic connections between various correlated indicators are uncovered, decomposing them into K groups of state linkage indicators. Subsequently, principal component analysis is used to carefully construct K surrounding rock deformation identification models from these groups of indicators, enabling accurate judgment of the surrounding rock deformation status based on input data. Finally, by connecting these models in parallel, a fully functional and efficient surrounding rock deformation monitoring cloud platform is constructed. Simultaneously, establishing a communication link between the monitoring array and the cloud platform is equally crucial. Based on the mapping relationship between multiple surrounding rock state correlation indicators and the K groups of state linkage indicators, numerous surrounding rock state monitoring devices are rationally divided into K groups. Each group of devices is equipped with a dedicated data relay edge node, and a stable and reliable wired communication link is constructed to ensure high-speed and stable transmission of monitoring data from the device end to the edge node. Next, a wireless communication link is established between the edge node and the K surrounding rock deformation identification models in the cloud. High-speed wireless network technology is used to realize the rapid uploading of data and the timely issuance of instructions, enabling the entire monitoring system to achieve real-time and efficient data interaction and ensuring the continuity and accuracy of monitoring work.

[0025] Step S400: When the surrounding rock deformation monitoring cloud receives and analyzes the data transmitted back from the surrounding rock deformation monitoring array and outputs a surrounding rock deformation warning, a source tracing analysis is performed based on the surrounding rock deformation warning to obtain the predicted defects in construction technology.

[0026] Specifically, after the surrounding rock deformation monitoring cloud receives data transmitted back from the monitoring array, it first uses the Support Vector Machine (SVM) algorithm to label different degrees of surrounding rock deformation in historical monitoring data into different categories. By finding the optimal hyperplane to distinguish each category of data, a classification model is constructed to classify new data. Simultaneously, a decision tree algorithm is used to construct a tree structure based on features such as displacement rate and strain change, and the deformation category of new data is determined by following the branches according to feature values. For deep learning technology, a Convolutional Neural Network (CNN) is used to convert continuous time-series displacement and strain data into image form. Spatiotemporal feature patterns are automatically extracted through convolutional layers, pooling layers, and fully connected layers to determine the surrounding rock deformation trend. Furthermore, by utilizing recurrent neural networks (RNNs) and their variants (such as LSTM and GRU), leveraging their memory function and special gating mechanisms, sequential data is processed to capture temporal dependencies. By integrating the analysis results of these algorithms and technologies, once a rock deformation early warning is output, the outlier values ​​in the warning are input based on the construction deformation causal network. Through the relationships between nodes and causal edges, the causal path of construction deformation is identified, relevant construction technologies and probabilities are extracted, and risks are assessed in conjunction with historical construction time, ultimately determining the defects in the construction technology prediction.

[0027] Step S500: Based on the construction technology, predict the defects and locate the construction deviation to obtain the construction deviation prediction result.

[0028] Specifically, in the quality and safety management process of tunnel construction, the aim is to accurately locate construction deviations to ensure smooth construction progress. Once predicted defects in construction techniques are obtained, a professional technical team will conduct in-depth work to pinpoint these deviations. For each construction technique involved in the predicted defects, such as dewatering testing and advanced support techniques, detailed construction records will be reviewed, including parameter settings, construction sequence, and material usage. For example, if advanced support techniques are determined to have defects, the length, spacing, and anchoring force of anchor bolts will be carefully checked to ensure they meet design requirements, as well as the installation angle of the support structure and the thickness and strength of the shotcrete. Simultaneously, a comprehensive analysis will be conducted considering the actual conditions at the construction site, such as changes in geological conditions and the influence of the construction environment. By comparing the design scheme with the actual construction operation, differences between the two will be identified; these differences represent potential construction deviations. Advanced measuring instruments and testing equipment will be used to accurately measure the actual state of the tunnel surrounding rock again, comparing and verifying the data with previously collected data from the monitoring array to further clarify the specific location and extent of the deviations. Furthermore, the analysis will reference construction deviations and solutions observed in similar engineering cases to conduct a more comprehensive and in-depth analysis of current construction deviations. Through this rigorous and meticulous process, the predicted construction deviations are ultimately obtained, providing an accurate basis for developing targeted corrective measures and ensuring the safe and efficient progress of tunnel construction.

[0029] Step S600: Generate a construction early warning instruction based on the construction deviation prediction result, and send the construction early warning instruction to the construction operation and maintenance center.

[0030] Specifically, based on the obtained accurate construction deviation prediction results, a construction early warning instruction is quickly generated. This instruction is comprehensive and highly targeted, first clearly describing the specific circumstances of the construction deviation, including its type, location, degree, and potential risks, such as further deterioration of surrounding rock stability and increased risk of water and sand inrush. Simultaneously, the instruction provides detailed rectification suggestions, such as adjusting construction parameters, optimizing construction techniques, and replacing construction materials, along with the implementation priority and estimated completion time for each measure. After generating the instruction, it is sent to the construction operation and maintenance center as quickly as possible using an efficient and reliable communication system, such as a dedicated wireless network, wired communication lines, or a comprehensive communication platform. Strict adherence to the rectification suggestions ensures timely correction of deviations, restoration of normal construction order, effective protection of tunnel construction safety, improved construction efficiency, and reduction of project delays and safety risks caused by construction deviations.

[0031] In one possible implementation, step S300 further includes:

[0032] Step S310: By performing index linkage effect analysis on the surrounding rock state correlation set, the surrounding rock state correlation set is decomposed into K sets of state linkage indices.

[0033] Step S320: By performing principal component analysis on the K groups of state linkage indices, K surrounding rock deformation identification models are constructed.

[0034] Step S330: By connecting the K surrounding rock deformation identification models in parallel, the construction of the surrounding rock deformation monitoring cloud platform is completed.

[0035] Specifically, in the initial stage of constructing the cloud-based monitoring system for surrounding rock deformation, a detailed analysis of the linkage effects of various indicators is conducted to uncover the intrinsic connections between them, decomposing them into K sets of state-linked indicators. Taking the linkage indicator set of displacement, stress, and crack propagation as an example, in actual tunnel construction scenarios, if the displacement of the surrounding rock continues to increase, it may be due to changes in internal stress. When the stress exceeds the safe range, the internal structure of the surrounding rock cannot withstand it, which will lead to the generation and propagation of cracks. There is a close causal relationship between this displacement, stress, and crack propagation; they influence and interact with each other, jointly reflecting the deformation state of the surrounding rock. Once the displacement continues to increase and the stress exceeds the standard, accompanied by the continuous propagation of cracks, it often indicates that the deformation and damage of the surrounding rock has approached an irreversible and dangerous stage. Next, consider the linkage indicator set of pore water pressure, displacement, and ground settlement. In complex and weak strata, changes in pore water pressure have a significant impact on the stability of the surrounding rock. When the pore water pressure increases, it will generate additional forces on the surrounding rock, changing the stress state of the surrounding rock and leading to increased displacement. The changes in surrounding rock displacement will then be transmitted to the surface, causing ground settlement. Especially when the water flow direction is unreasonable or the precipitation measures are inadequate, a vicious cycle forms between pore water pressure and surrounding rock deformation. Water pressure promotes surrounding rock deformation, which in turn affects the water flow channel, further altering the water pressure distribution, and may ultimately lead to severe surface subsidence and surrounding rock failure. Through precise analysis of these complex interrelationships, the surrounding rock state correlation set is scientifically decomposed into K sets of state linkage indicators, laying the foundation for the subsequent construction of a precise monitoring model.

[0036] After grouping the indicators, principal component analysis (PCA) is performed on the K groups of state-linked indicators to construct K surrounding rock deformation identification models. PCA aims to extract key information from numerous related variables (i.e., state-linked indicators) and replace the original indicators with a few unrelated comprehensive indicators (principal components), thereby simplifying the data structure while retaining the main characteristics of the original data. For each group of state-linked indicators, PCA assigns different weights based on the importance of each indicator's impact on surrounding rock deformation, transforming it into a new set of comprehensive indicators, i.e., principal components. These principal components can reflect the information contained in the original indicators to the greatest extent and are independent of each other. For example, when analyzing a group of state-linked indicators including displacement rate of change, stress gradient, and crack width change, PCA may find that the displacement rate of change plays a dominant role in reflecting the current deformation trend of the surrounding rock, the stress gradient has a significant impact on the potential development of deformation, and crack width change is a key factor in judging the degree of damage to the surrounding rock structure. By reasonably weighting and combining these indicators, a principal component that can comprehensively characterize the features of this group of indicators is formed. By processing the K groups of state-linked indicators in this way, K principal components are obtained. Each principal component summarizes the relationship between the corresponding group of indicators and the surrounding rock deformation from different perspectives. Based on these principal components, K surrounding rock deformation identification models are constructed, which can accurately determine the deformation state of the surrounding rock based on the input monitoring data, providing strong support for subsequent cloud-based decision-making.

[0037] The construction of the cloud-based rock deformation monitoring system has been completed. Parallel operation means that these models run simultaneously on the cloud platform, jointly processing real-time data from the monitoring array. Each model independently analyzes and judges the data, then aggregates its results to the integrated decision-making module in the cloud. This parallel structure fully leverages the strengths of each model, improving the system's reliability and accuracy. When monitoring data is transmitted to the cloud, K models receive and process it simultaneously. For example, one model may have unique advantages in analyzing displacement-related data, while another model is more sensitive to changes in pore water pressure; they each interpret the data from different dimensions. The integrated decision-making module integrates the outputs of each model and, through weighted averaging, majority voting, or other fusion algorithms, arrives at the final judgment on rock deformation. This cloud-based rock deformation monitoring system can process massive amounts of monitoring data in real time and efficiently, accurately and promptly judging the rock deformation situation. Once an anomaly is detected, it can quickly issue early warning signals, providing strong protection for construction safety and ensuring the safe and orderly progress of tunnel construction under complex geological conditions.

[0038] In one possible implementation, step S320 further includes:

[0039] Step S321: Call the surrounding rock deformation data of the K group of state linkage indicators to obtain the K group of historical state anomaly sequences.

[0040] Step S322: Statistical analysis of the recurrence frequency of the K-group state linkage indicators in the K-group historical state anomaly sequences to obtain the K-group sequence location information.

[0041] Step S323: Configure the linkage hierarchy of the K groups of state linkage indicators based on the K groups of sequence position information to obtain K state anomaly monitoring sequences.

[0042] Step S324: Preconstruct a set of surrounding rock risk identification models corresponding to the surrounding rock state association set.

[0043] Step S325: Based on the K abnormal state monitoring sequences, perform model invocation and hierarchical connection in the surrounding rock risk identification model set to complete the construction of the K surrounding rock deformation identification models.

[0044] Specifically, in the process of constructing the surrounding rock deformation identification model, for the K sets of state linkage indicators, it is necessary to deeply explore historical monitoring data resources and retrieve surrounding rock deformation data. This historical data covers monitoring information from different stages and geological conditions during past tunnel construction, providing valuable evidence for analyzing the deformation patterns of the surrounding rock. By screening and organizing massive amounts of historical data, and using each set of state linkage indicators as a clue, data segments showing abnormal states are identified, thereby obtaining the K sets of historical state anomaly sequences. For example, for the displacement and stress linkage indicators, when historical data shows a sharp increase in displacement and stress far exceeding the normal range within a certain period, these consecutive data points constitute a set of historical state anomaly sequences. These anomaly sequences reflect the instability situations that the surrounding rock may have faced during past construction, providing key samples for subsequent analysis. They help identify under what combination of indicator changes the surrounding rock is prone to problems, thus accumulating empirical data for constructing an accurate deformation identification model.

[0045] To statistically analyze the recurrence frequency of K sets of state-linked indicators and obtain the positional information of K sets of sequences in K sets of historical state anomaly sequences, a sliding window algorithm combined with a hash table data structure is employed. First, a sliding window of appropriate size is set for each set of historical state anomaly sequences, sliding from the beginning of the sequence. For each state-linked indicator value within the window, a specific hash function is used to map it to a unique hash value, which is then stored in the hash table. Simultaneously, the position of the window within the sequence is recorded as the sequence position information. After each sliding window operation, the occurrence count of the corresponding hash value in the hash table is updated, thus achieving the statistical analysis of the indicator sequence occurrence frequency. After traversing the entire historical state anomaly sequence, the hash table stores the recurrence frequency of each indicator sequence and its corresponding sequence position information, providing crucial data support for subsequent configuration of linkage hierarchy relationships. This facilitates a more accurate construction of the surrounding rock deformation identification model, ensuring tunnel construction safety.

[0046] To configure the hierarchical relationship of K sets of state linkage indicators based on K sets of sequence location information and obtain K state anomaly monitoring sequences, an algorithm based on time series analysis and causal relationship mining is adopted. First, the time series data corresponding to each set of sequence location information is preprocessed, such as through data cleaning and interpolation, to ensure data integrity and accuracy. Then, the Granger causality test is used to analyze the causal relationships between state linkage indicators. If indicator A changes before indicator B in time, and the change in indicator A significantly affects the change trend of indicator B, then indicator A is determined to have a causal relationship with indicator B, and indicator A is placed at a higher level. Based on this causal relationship analysis result, a directed acyclic graph (DAG) is constructed to represent the linkage hierarchy. Finally, based on the constructed DAG, starting from the initial node (i.e., the indicator at the highest level), indicators are combined sequentially according to the hierarchy to form K state anomaly monitoring sequences, thereby providing strong support for accurately monitoring the deformation state of the surrounding rock and ensuring tunnel construction safety.

[0047] When pre-constructing the set of surrounding rock risk identification models corresponding to the surrounding rock state association set, the random forest algorithm is used to construct multiple decision tree models by sampling with replacement from a large amount of historical construction data (including surrounding rock state parameters, construction technology information, and risk events that have occurred). Each decision tree model evaluates the surrounding rock risk from different perspectives. These decision trees together form the set of surrounding rock risk identification models, which comprehensively considers the impact of multiple factors on the stability of the surrounding rock.

[0048] Based on K anomaly monitoring sequences, when calling and hierarchically connecting models in a rock risk identification model set, a Bayesian network algorithm is used to construct the connection relationships between models. For each anomaly monitoring sequence, the conditional probability relationships between indicators are analyzed and transformed into nodes and directed edges in a Bayesian network to determine the dependencies between models. Then, according to the structure of the Bayesian network, starting from the model related to the initial indicator of the sequence, subsequent related models are called and connected sequentially along the directed edges. The parameters of each model are updated through Bayesian inference, ultimately completing the construction of K rock deformation identification models, achieving accurate prediction of rock deformation states, and ensuring tunnel construction safety.

[0049] In one possible implementation, step S300 further includes:

[0050] Step S340: Based on the mapping relationship between the multiple surrounding rock condition correlation indicators and the K group of condition linkage indicators, the multiple surrounding rock condition monitoring devices are divided into K groups of surrounding rock condition monitoring devices.

[0051] Step S350: Configure K data relay edge nodes for the K groups of surrounding rock condition monitoring equipment.

[0052] Step S360: Construct wired communication links between the K groups of surrounding rock condition monitoring devices and the K data relay edge nodes.

[0053] Step S370: Construct wireless communication links between the K data relay edge nodes and the K surrounding rock deformation identification models in the surrounding rock deformation monitoring cloud.

[0054] Specifically, based on the established mapping relationship between multiple surrounding rock condition correlation indicators and K sets of condition linkage indicators, numerous surrounding rock condition monitoring devices are scientifically grouped. Each surrounding rock condition correlation indicator corresponds to a specific monitoring task, and these indicators have an inherent logical connection with the K sets of condition linkage indicators. For example, displacement indicators may be closely related to a set of condition linkage indicators including displacement, stress, and crack propagation, because displacement changes are often accompanied by stress adjustments and crack development. According to this mapping relationship, sensors and other devices responsible for monitoring displacement are assigned to the corresponding group of surrounding rock condition monitoring devices. Similarly, devices monitoring pore water pressure are grouped according to their relationship with the set of condition linkage indicators of pore water pressure, displacement, and ground settlement. Through this detailed division, each group of surrounding rock condition monitoring devices can work closely around a set of condition linkage indicators, ensuring the accuracy and effectiveness of subsequent data acquisition and transmission. This ensures that the monitoring system can comprehensively and accurately capture information on changes in surrounding rock condition from multiple dimensions, providing reliable data support for tunnel construction safety.

[0055] After grouping the monitoring equipment, the focus shifted to configuring K data relay edge nodes for each of the K groups of surrounding rock condition monitoring equipment. Each edge node possesses certain data processing capabilities and storage capacity, enabling it to perform preliminary processing on data from the connected group of surrounding rock condition monitoring equipment, such as data cleaning and format conversion. This reduces the burden on cloud data processing and improves data transmission efficiency. For example, for the raw displacement data detected, the edge node can perform simple data verification locally, removing obviously abnormal data points, and then package the processed data for transmission. Simultaneously, the edge node configuration makes the connection between the monitoring equipment and the cloud more flexible and stable. Even in complex tunnel construction environments, such as those with electromagnetic interference or signal obstruction, it ensures a reliable data relay channel, guaranteeing data transmission continuity. This lays a solid foundation for building an efficient communication link, effectively preventing monitoring vulnerabilities caused by data transmission interruptions, and further ensuring tunnel construction safety.

[0056] After configuring the edge nodes, wired communication links are established between the K groups of surrounding rock condition monitoring devices and the K data relay edge nodes. These wired communication links, with their high stability and reliability, are crucial for ensuring stable data transmission. For each group of surrounding rock condition monitoring devices and its corresponding edge nodes, appropriate wired communication technologies, such as Ethernet and RS-485 standard bus technologies, are used for connection. These technologies feature high transmission speeds and strong anti-interference capabilities, ensuring the accurate transmission of massive amounts of data collected by the monitoring devices to the edge nodes. For example, in tunnel construction environments, despite electromagnetic interference from various mechanical equipment and complex geological conditions, the wired communication links effectively shield these interferences, ensuring the stable transmission of surrounding rock condition data collected in real-time by monitoring devices such as displacement sensors and stress gauges to the edge nodes. Furthermore, the wired communication links are easy to maintain and manage; any faults can be quickly located and repaired, ensuring that the data acquisition and transmission of the entire monitoring system are unaffected. This provides continuous and reliable data support for tunnel construction safety, enabling the construction team to promptly grasp changes in the surrounding rock condition and make accurate construction decisions.

[0057] As the final step in building the communication links, a wireless communication link is constructed between K data relay edge nodes and K surrounding rock deformation identification models in the cloud-based surrounding rock deformation monitoring system. This wireless communication link injects flexibility and efficiency into the entire monitoring system, enabling rapid data upload and timely command issuance. Advanced wireless communication technologies, such as 5G and Wi-Fi 6, are employed to establish a high-speed data channel between the data relay edge nodes and the cloud. These technologies meet the monitoring system's requirements for large-volume data transmission and low latency, ensuring that data processed by the edge nodes can be rapidly uploaded to the cloud for real-time analysis by the K surrounding rock deformation identification models. Simultaneously, commands generated by the cloud based on the analysis results can be quickly issued to the edge nodes via the wireless communication link, and then transmitted to the corresponding monitoring equipment, enabling real-time control and adjustment of the tunnel construction process. For example, when the cloud model analysis detects an abnormal trend in surrounding rock deformation, it immediately sends an early warning command to the edge nodes via the wireless communication link. The edge nodes then notify the relevant monitoring equipment to initiate a more intensive data acquisition mode, allowing the construction team to take timely measures, such as strengthening support or adjusting construction techniques, effectively ensuring tunnel construction safety, avoiding safety accidents caused by communication delays, and ensuring smooth construction progress.

[0058] In one possible implementation, step S400 further includes:

[0059] Step S410: The surrounding rock deformation monitoring cloud receives the real-time monitoring data array transmitted back by the surrounding rock deformation monitoring array.

[0060] Step S420: Based on the mapping relationship between the multiple surrounding rock condition monitoring devices and the multiple surrounding rock condition correlation indicators, and with the K groups of surrounding rock condition monitoring devices as constraints, the real-time monitoring data array is divided into K groups of real-time monitoring data.

[0061] Step S430: After loading and sorting the K sets of real-time monitoring data according to the K state anomaly monitoring sequences, load the K sets of real-time monitoring data into the K surrounding rock deformation identification models to indirectly judge the surrounding rock deformation, and output the K surrounding rock deformation judgment results.

[0062] Step S440: When any of the K surrounding rock deformation judgment results is a non-empty set, the surrounding rock deformation early warning is generated, wherein the surrounding rock deformation early warning includes H surrounding rock state associated anomalies.

[0063] Step S450: Based on the surrounding rock deformation early warning, perform source tracing analysis to obtain the predicted defects of the construction technology.

[0064] Specifically, in the key process of tunnel construction safety monitoring, a crucial stage of data processing and analysis has begun. The cloud-based rock deformation monitoring system receives real-time monitoring data from the surrounding rock deformation monitoring array. This real-time monitoring data contains rich information about the state of the surrounding rock within the tunnel construction area, collected continuously by multiple surrounding rock condition monitoring devices distributed at various key locations. The data is transmitted to the cloud at high speed and stably through a pre-built communication link. The cloud has powerful data receiving capabilities, capable of simultaneously processing data traffic from numerous monitoring devices, ensuring data integrity and timeliness. This provides sufficient raw data for subsequent accurate analysis, laying the foundation for timely understanding of dynamic changes in the surrounding rock, and serving as the first line of defense for ensuring tunnel construction safety.

[0065] Once the cloud successfully receives the real-time monitoring data array, it scientifically divides the massive data array based on a carefully established mapping relationship between multiple surrounding rock condition correlation indicators and multiple surrounding rock condition monitoring devices, using K groups of surrounding rock condition monitoring devices as constraints. For example, data collected by monitoring devices corresponding to displacement indicators will be classified into the group of real-time monitoring data related to displacement. Similarly, data on indicators such as stress and pore water pressure will be accurately categorized according to their corresponding monitoring devices. This classification method allows the data to be organized in an orderly manner according to the surrounding rock condition characteristics it reflects, facilitating subsequent targeted analysis of different groups of data. This ensures that each data point can play its maximum value in the corresponding analysis model, providing a refined data foundation for accurately judging the deformation state of the surrounding rock and further improving the level of construction safety assurance.

[0066] After data segmentation, K sets of real-time monitoring data were carefully loaded and sorted according to K anomaly monitoring sequences to ensure the input order conformed to the pre-defined monitoring logic. The sorted K sets of real-time monitoring data were then sequentially loaded into K surrounding rock deformation identification models. These models utilize advanced algorithms and deep learning techniques to deeply mine and analyze the data. Through learning from historical data and model training, the models can identify characteristic patterns in the data, thereby indirectly judging the surrounding rock deformation situation. For example, the model can comprehensively judge whether the current deformation state of the surrounding rock is within the normal range based on multiple factors such as displacement data trends, stress distribution, and crack propagation characteristics, ultimately outputting K surrounding rock deformation judgment results. This provides crucial evidence for timely detection of potential surrounding rock deformation risks, ensuring that construction safety hazards can be detected and addressed as early as possible.

[0067] After obtaining K surrounding rock deformation assessment results, the results are evaluated and an early warning is issued. If any of the K assessment results is a non-empty set, it indicates an abnormal signal in the monitoring data, suggesting that the surrounding rock deformation exceeds the normal range. At this point, a surrounding rock deformation early warning is immediately generated. This warning includes H surrounding rock state-related anomaly values, i.e., surrounding rock state-related indicators. These anomaly values ​​are key clues for judging abnormal surrounding rock conditions, involving multiple aspects such as sudden increases in displacement, abnormal stress concentration, and rapid changes in pore water pressure. The early warning information will be promptly transmitted to relevant personnel so that they can quickly take countermeasures, such as strengthening on-site monitoring and organizing expert consultations, to ensure timely and effective protection of tunnel construction safety and avoid safety accidents caused by surrounding rock deformation.

[0068] After generating a rock deformation early warning, in-depth source analysis is used to find the root cause of the problem, thereby obtaining the predicted defects in construction technology. Based on the previously constructed causal network of construction deformation, H outliers from the rock deformation early warning are input into this network. The causal network of construction deformation is constructed by transforming multiple rock state correlation indicators into state-related causal nodes, abstracting various sample construction technologies into causal nodes of construction defects, constructing causal edges based on rich historical construction defect data, and then rationally allocating conditional probabilities according to the frequency of overlap of causal edges. By searching for paths related to outliers in this complex network, H causal paths of construction deformation are accurately identified, and then H groups of related construction technologies and their corresponding H groups of causal conditional probabilities are extracted. Further interactive acquisition of the historical construction time of these construction technologies is conducted, and a comprehensive risk contribution evaluation is performed using a risk assessment model, taking into account factors such as causal conditional probabilities and historical construction time. Finally, W construction technologies with risk contributions exceeding a preset threshold are accurately identified as predicted defects in construction technology, providing a clear direction for subsequent targeted improvement of construction processes and adjustment of construction plans, ensuring that tunnel construction can be carried out safely and efficiently.

[0069] In one possible implementation, step S440 further includes:

[0070] Step S441: Construct multiple state-related causal nodes based on the multiple surrounding rock state correlation indicators.

[0071] Step S442: Interact to obtain multiple sample construction techniques, and construct multiple causal nodes of construction defects based on the multiple sample construction techniques.

[0072] Step S443: Interact to obtain historical construction defect data, and construct causal edges between the multiple state-related causal nodes and the multiple construction defect causal nodes based on the historical construction defect data.

[0073] Step S444: Configure conditional probability allocation based on the frequency of overlap of causal edges between causal nodes to complete the construction of the causal network for construction deformation.

[0074] Specifically, multiple state-related causal nodes are constructed based on various surrounding rock condition correlation indicators. These indicators are key elements comprehensively reflecting the surrounding rock condition, such as displacement, stress, crack propagation, and pore water pressure. Each indicator is transformed into a state-related causal node to capture information on changes in the surrounding rock condition. Taking displacement as an example, the state-related causal node corresponding to the displacement indicator can reflect the changes in the spatial position of the surrounding rock in real time. Whether it is a small displacement adjustment or a large-scale movement trend, it can be characterized through this node. These nodes are interconnected yet independent, collectively forming the basic framework of the construction deformation causal network. This provides a crucial information carrier for subsequent accurate analysis of the causal relationship between surrounding rock deformation and construction technology, and is the cornerstone for constructing a precise causal network.

[0075] After constructing the state-related causal nodes, the focus is on acquiring various sample construction technologies through interaction, and then constructing multiple construction defect causal nodes based on these sample construction technologies. The sample construction technologies cover various processes, methods, and operational procedures involved in tunnel construction, such as dewatering testing, advanced support, and lining construction. For each sample construction technology, potential defects during construction are analyzed, and these potential defects are abstracted into construction defect causal nodes. For example, in advanced support technology, if the anchor bolt installation angle is inaccurate or the anchoring force is insufficient, these potential defects are defined as corresponding construction defect causal nodes. These nodes represent weak points in the construction process, and they interact with the state-related causal nodes to form a complex causal relationship network. This provides important node support for in-depth exploration of the intrinsic connection between construction technology and surrounding rock deformation, and helps to accurately trace the root cause of construction problems.

[0076] After the initial construction of state-related causal nodes and construction defect causal nodes, historical construction defect data is interactively obtained. Based on this rich data, causal edges are constructed between multiple state-related causal nodes and multiple construction defect causal nodes. Historical construction defect data records various problems that actually occurred during past tunnel construction, as well as the surrounding rock conditions and construction techniques at the time. Through in-depth analysis of this data, if a specific construction defect (such as insufficient dewatering leading to abnormal pore water pressure) is found to be consistently accompanied by specific changes in the surrounding rock condition (such as increased surrounding rock displacement), then a causal edge is established between the corresponding construction defect causal node and state-related causal node. The direction of the causal edge indicates the direction of the causal relationship, i.e., from the construction defect to the change in the surrounding rock condition. The existence or absence of causal edges and the combination of nodes they connect intuitively reflect the correlation pattern between construction defects and surrounding rock deformation, providing a structural foundation for subsequent probabilistic analysis and enabling the identification of clear causal relationships from complex data.

[0077] As the final crucial step in constructing the causal network of construction deformation, conditional probability allocation is configured based on the overlap frequency of causal edges between causal nodes, thus completing the construction of the entire network. The overlap frequency of causal edges is an important criterion for measuring how a construction defect leads to an abnormal state of a specific surrounding rock. If a causal edge appears frequently in historical data, it indicates a strong causal relationship between the construction defect and the corresponding change in the surrounding rock state. Based on this, a reasonable conditional probability value is assigned to each causal edge, representing the probability of an abnormal surrounding rock state occurring under the condition of a specific construction defect. For example, if historical data shows that the probability of rock fissure propagation is high when insufficient anchor bolt anchoring force occurs multiple times in advanced support technology, then a higher conditional probability value will be assigned to the causal edge connecting the causal node of insufficient anchor bolt anchoring force (the construction defect) in advanced support technology with the causal node related to the state of rock fissure propagation. Through such detailed configuration of conditional probabilities, the causal network of construction deformation can be fully constructed, enabling it not only to qualitatively describe the causal relationship between construction technology and surrounding rock deformation, but also to quantitatively assess the risk probability of surrounding rock deformation under various conditions. This provides powerful tool support for subsequent accurate source tracing analysis and construction technology defect prediction, effectively ensuring tunnel construction safety.

[0078] In one possible implementation, step S450 further includes:

[0079] Step S451: Input the H surrounding rock state associated anomalies in the surrounding rock deformation early warning into the construction deformation causal network, and filter to obtain H construction deformation causal paths.

[0080] Step S452: Extract construction techniques based on the H causal paths of construction deformation to obtain H sets of construction techniques and H sets of causal condition probabilities.

[0081] Step S453: Interact to obtain H historical construction times of the H group of construction technologies.

[0082] Step S454: Evaluate the risk contribution of the H groups of construction technologies based on the H groups of causal conditional probabilities and H historical construction times, and obtain W construction technologies whose risk contribution is higher than a preset threshold, as the predicted defects of the construction technologies.

[0083] Specifically, to input H outlier values ​​associated with the surrounding rock conditions into the construction deformation causal network and select H causal paths for construction deformation, a depth-first search (DFS) algorithm is employed. Starting from the causal node associated with each outlier value, a depth-first traversal is performed along the causal edges. During the traversal, the nodes and edges visited are recorded, forming paths. When a causal node associated with a construction defect is encountered or the path cannot be further expanded, it is determined whether the path satisfies the conditions related to the outlier value (e.g., the combination of nodes and edges on the path can explain the cause of the outlier value). If satisfied, the path is saved as a causal path for construction deformation. This process is repeated until all paths originating from nodes corresponding to outliers have been searched, ultimately obtaining H causal paths for construction deformation. This provides crucial clues for accurately tracing the root causes of construction problems and ensuring tunnel construction safety.

[0084] For each causal path of construction deformation, a path parsing algorithm is used to extract construction techniques and obtain causal conditional probabilities. First, along the path from the starting node to the ending node, nodes representing construction techniques are identified, and their corresponding construction techniques are extracted, forming H groups of construction techniques. Then, causal edges connecting construction technique nodes and surrounding rock state-related nodes are found in the path, and pre-defined conditional probability values ​​are obtained for each causal edge. These values ​​are derived from statistical analysis of historical construction defect data. These conditional probability values ​​are grouped according to their corresponding construction techniques to obtain H groups of causal conditional probabilities. This approach provides an important data foundation for subsequent comprehensive assessment of construction technique risks, helps accurately determine the potential impact of different construction techniques on surrounding rock deformation, and ensures tunnel construction safety.

[0085] To interactively obtain H historical construction times for H groups of construction technologies, a database query and association algorithm is employed. First, the system connects to a database storing historical construction data, which covers detailed information about past tunnel construction projects, including records of construction technology applications and construction timelines. For each group of construction technologies, key features such as technology name and process parameters are used as query conditions to perform an exact match query in the database tables. By constructing association statements, the construction technology table and the time record table are joined internally according to association fields such as project number and construction stage, ensuring that the query results accurately correspond to the application period of each group of construction technologies. For example, if a group of construction technologies involves specific anchor bolt support parameters, the query statement precisely locates the construction time period using those parameters, extracts and associates them with the corresponding group, ultimately successfully obtaining H historical construction times interactively. This provides crucial time dimension information for subsequent accurate risk assessment, ensuring tunnel construction safety.

[0086] The risk contribution of H groups of construction technologies is evaluated based on H groups of causal conditional probabilities and H historical construction times, thereby identifying construction technology prediction defects. First, the Analytic Hierarchy Process (AHP) is used, leveraging the rich practical experience of experts and a large amount of historical data to construct a judgment matrix. Here, the H groups of causal conditional probabilities are used as one important evaluation indicator dimension, and the H historical construction times as another. For these two dimensions, their relative importance is carefully compared pairwise, and scores are assigned according to certain standards. Through rigorous calculation, the weight of each indicator is determined. Next, the fuzzy comprehensive evaluation method is introduced. For each group of construction technologies, its corresponding causal conditional probability value and the quantified value of its matching historical construction time are input as basic data. Based on the weights, these data are processed under specific fuzzy transformation rules to obtain a comprehensive evaluation vector. Finally, this comprehensive evaluation vector is compared one by one with pre-set thresholds. Construction technologies with comprehensive evaluation values ​​higher than the thresholds are the key risk points to be identified and marked as construction technology prediction defects. This allows for the precise identification of potential technical hazards affecting construction safety, providing crucial guidance for subsequent adjustments to the construction plan and ensuring the safe and smooth progress of tunnel construction.

[0087] Example 2, based on the same inventive concept as the tunnel surrounding rock deformation monitoring method in the foregoing examples, such as... Figure 2 As shown, this application provides a tunnel surrounding rock deformation monitoring system. The system and method embodiments in this application are based on the same inventive concept. The system includes:

[0088] The surrounding rock state association set acquisition module 10 is used to interactively obtain the surrounding rock state association set, wherein the surrounding rock state association set includes multiple surrounding rock state association indicators.

[0089] The surrounding rock deformation monitoring array generation module 20 is used to configure monitoring equipment in the tunnel construction area according to the surrounding rock state association set and generate a surrounding rock deformation monitoring array, wherein the surrounding rock deformation monitoring array includes multiple surrounding rock state monitoring devices corresponding to the multiple surrounding rock state association indicators.

[0090] The communication link construction module 30 is used to pre-build the surrounding rock deformation monitoring cloud and construct the communication link between the surrounding rock deformation monitoring array and the surrounding rock deformation monitoring cloud.

[0091] The construction technology prediction defect acquisition module 40 is used to perform source analysis based on the surrounding rock deformation warning when the surrounding rock deformation monitoring cloud receives and analyzes the feedback data of the surrounding rock deformation monitoring array and outputs a surrounding rock deformation warning, thereby obtaining the construction technology prediction defect.

[0092] The construction deviation prediction result acquisition module 50 is used to locate construction deviations based on the predicted defects of the construction technology and obtain construction deviation prediction results.

[0093] The construction early warning instruction generation module 60 is used to generate construction early warning instructions based on the construction deviation prediction results and send the construction early warning instructions to the construction operation and maintenance center.

[0094] Furthermore, the communication link construction module 30 also includes:

[0095] The index linkage effect analysis unit is used to decompose the surrounding rock state association set into K groups of state linkage indices by performing index linkage effect analysis on the surrounding rock state association set.

[0096] The surrounding rock deformation identification model construction unit is used to construct K surrounding rock deformation identification models by performing principal component analysis on the K groups of state linkage indicators.

[0097] The surrounding rock deformation monitoring cloud construction unit is used to complete the construction of the surrounding rock deformation monitoring cloud by connecting the K surrounding rock deformation identification models in parallel.

[0098] Furthermore, the surrounding rock deformation identification model construction unit also includes:

[0099] The historical state anomaly sequence acquisition unit is used to retrieve surrounding rock deformation data from the K sets of state linkage indicators to obtain the K sets of historical state anomaly sequences.

[0100] The sequence location information acquisition unit is used to perform frequency statistics of the index sequence recurrence of the K groups of state linkage indicators in the K groups of historical state abnormal sequences, and obtain the sequence location information of the K groups.

[0101] The status anomaly monitoring sequence acquisition unit is used to configure the linkage hierarchy relationship of the K groups of status linkage indicators based on the K groups of sequence position information to obtain K status anomaly monitoring sequences.

[0102] A pre-construction unit for the surrounding rock risk identification model set is used to pre-construct a surrounding rock risk identification model set corresponding to the surrounding rock state association set.

[0103] The hierarchical connection unit is used to call and connect the K state anomaly monitoring sequences in the surrounding rock risk identification model set to complete the construction of the K surrounding rock deformation identification models.

[0104] Furthermore, the communication link construction module 30 also includes:

[0105] The surrounding rock condition monitoring equipment division unit is used to divide the multiple surrounding rock condition monitoring devices into K groups of surrounding rock condition monitoring devices according to the mapping relationship between the multiple surrounding rock condition correlation indicators and K groups of condition linkage indicators.

[0106] The data relay edge node configuration unit is used to configure K data relay edge nodes for the K groups of surrounding rock condition monitoring devices.

[0107] A wired communication link construction unit is used to construct wired communication links between the K groups of surrounding rock condition monitoring devices and the K data relay edge nodes.

[0108] The wireless communication link construction unit is used to construct wireless communication links between the K data relay edge nodes and the K surrounding rock deformation identification models in the surrounding rock deformation monitoring cloud.

[0109] Furthermore, the construction technology prediction defect acquisition module 40 also includes:

[0110] The real-time monitoring data array receiving unit is used to receive the real-time monitoring data array transmitted back by the surrounding rock deformation monitoring array from the surrounding rock deformation monitoring cloud.

[0111] The real-time monitoring data partitioning unit is used to partition the real-time monitoring data array into K groups of real-time monitoring data based on the mapping relationship between the multiple surrounding rock condition correlation indicators and the multiple surrounding rock condition monitoring devices, with the K groups of surrounding rock condition monitoring devices as constraints.

[0112] The surrounding rock deformation judgment result output unit is used to load and sort the K sets of real-time monitoring data according to the K state anomaly monitoring sequences, load the K sets of real-time monitoring data into the K surrounding rock deformation identification models to indirectly judge the surrounding rock deformation, and output K surrounding rock deformation judgment results.

[0113] The surrounding rock deformation early warning generation unit is used to generate the surrounding rock deformation early warning when any of the K surrounding rock deformation judgment results is a non-empty set, wherein the surrounding rock deformation early warning includes H surrounding rock state associated anomaly values.

[0114] The source analysis unit is used to perform source analysis based on the surrounding rock deformation early warning to obtain the predicted defects of the construction technology.

[0115] Furthermore, the surrounding rock deformation early warning generation unit also includes:

[0116] The state-related causal node construction unit constructs multiple state-related causal nodes based on the multiple surrounding rock state-related indicators.

[0117] The construction defect causal node construction unit is used to interactively obtain multiple sample construction technologies and construct multiple construction defect causal nodes based on the multiple sample construction technologies.

[0118] The causal edge construction unit is used to interactively obtain historical construction defect data and construct causal edges between the multiple state-related causal nodes and the multiple construction defect causal nodes based on the historical construction defect data.

[0119] The construction deformation causal network construction unit is used to configure conditional probability allocation based on the frequency of overlap of causal edges between causal nodes, thereby completing the construction of the construction deformation causal network.

[0120] Furthermore, the source tracing analysis unit also includes:

[0121] The construction deformation causal path acquisition unit is used to input the H surrounding rock state associated anomalies in the surrounding rock deformation early warning into the construction deformation causal network and filter to obtain H construction deformation causal paths.

[0122] The causal conditional probability acquisition unit is used to extract construction techniques based on the H causal paths of construction deformation, and obtain H sets of construction techniques and H sets of causal conditional probabilities.

[0123] The historical construction time acquisition unit is used to interactively obtain H historical construction times of the H groups of construction technologies.

[0124] The construction technology acquisition unit is used to evaluate the risk contribution of the H groups of construction technologies based on the H groups of causal condition probabilities and H historical construction times, and to obtain W construction technologies whose risk contribution is higher than a preset threshold, as the predicted defects of the construction technologies.

[0125] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0126] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0127] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A method for monitoring deformation of surrounding rock in tunnels, characterized in that, The method includes: Interactively obtain a set of surrounding rock state correlations, wherein the set of surrounding rock state correlations includes multiple surrounding rock state correlation indicators; Based on the surrounding rock state association set, monitoring equipment is configured in the tunnel construction area to generate a surrounding rock deformation monitoring array, wherein the surrounding rock deformation monitoring array includes multiple surrounding rock state monitoring devices corresponding to the multiple surrounding rock state association indicators; A cloud platform for monitoring surrounding rock deformation is pre-constructed, and a communication link is established between the surrounding rock deformation monitoring array and the cloud platform. When the surrounding rock deformation monitoring cloud receives and analyzes the data transmitted back from the surrounding rock deformation monitoring array, and outputs a surrounding rock deformation warning, a source tracing analysis is performed based on the surrounding rock deformation warning to obtain predicted defects in construction technology. Based on the construction technology, defects are predicted to locate construction deviations and the construction deviation prediction results are obtained. A construction early warning instruction is generated based on the construction deviation prediction results, and the construction early warning instruction is sent to the construction operation and maintenance center; A pre-built cloud platform for monitoring surrounding rock deformation includes: By analyzing the linkage effect of the indexes on the surrounding rock state correlation set, the surrounding rock state correlation set is decomposed into K sets of state linkage indices. By performing principal component analysis on the K sets of state linkage indices, K surrounding rock deformation identification models are constructed. By connecting the K surrounding rock deformation identification models in parallel, the cloud-based monitoring platform for surrounding rock deformation is constructed. By performing principal component analysis on the K sets of state linkage indices, K surrounding rock deformation identification models are constructed, including: The surrounding rock deformation data is retrieved from the K groups of state linkage indicators to obtain the K groups of historical state anomaly sequences; The frequency of recurrence of the index sequences of the K sets of historical abnormal state sequences is statistically analyzed to obtain the position information of the K sets of sequences. Based on the K sets of sequence position information, configure the linkage hierarchy of the K sets of state linkage indicators to obtain K state anomaly monitoring sequences; A set of surrounding rock risk identification models corresponding to the surrounding rock state association set is pre-constructed; Based on the K abnormal state monitoring sequences, model calls and hierarchical connections are performed in the surrounding rock risk identification model set to complete the construction of the K surrounding rock deformation identification models.

2. The method for monitoring tunnel surrounding rock deformation as described in claim 1, characterized in that, The method for establishing a communication link between the surrounding rock deformation monitoring array and the surrounding rock deformation monitoring cloud includes: Based on the mapping relationship between the multiple surrounding rock condition correlation indicators and the K group of condition linkage indicators, the multiple surrounding rock condition monitoring devices are divided into K groups of surrounding rock condition monitoring devices. Configure K data relay edge nodes for the K groups of surrounding rock condition monitoring equipment; Establish wired communication links between the K groups of surrounding rock condition monitoring devices and the K data relay edge nodes; A wireless communication link is constructed between the K data relay edge nodes and the K surrounding rock deformation identification models in the surrounding rock deformation monitoring cloud.

3. The method for monitoring tunnel surrounding rock deformation as described in claim 2, characterized in that, When the surrounding rock deformation monitoring cloud receives and analyzes the data transmitted back from the surrounding rock deformation monitoring array, and outputs a surrounding rock deformation early warning, a source tracing analysis is performed based on the surrounding rock deformation early warning to obtain predicted defects in construction technology. The method includes: The surrounding rock deformation monitoring cloud receives the real-time monitoring data array transmitted back by the surrounding rock deformation monitoring array; Based on the mapping relationship between the multiple surrounding rock condition correlation indicators and the multiple surrounding rock condition monitoring devices, and constrained by the K groups of surrounding rock condition monitoring devices, the real-time monitoring data array is divided into K groups of real-time monitoring data. After loading and sorting the K sets of real-time monitoring data according to the K state anomaly monitoring sequences, the K sets of real-time monitoring data are loaded into the K surrounding rock deformation identification models to indirectly judge the surrounding rock deformation, and K surrounding rock deformation judgment results are output. When any one of the K surrounding rock deformation judgment results is a non-empty set, the surrounding rock deformation early warning is generated, wherein the surrounding rock deformation early warning includes H surrounding rock state associated anomaly values; Based on the early warning of surrounding rock deformation, a source tracing analysis is performed to obtain the predicted defects of the construction technology.

4. The method for monitoring tunnel surrounding rock deformation as described in claim 3, characterized in that, The method includes: Multiple state-related causal nodes are constructed based on the aforementioned multiple surrounding rock state correlation indicators; Multiple sample construction techniques are obtained interactively, and multiple causal nodes of construction defects are constructed based on these multiple sample construction techniques; Historical construction defect data is obtained interactively, and causal edges between the multiple state-related causal nodes and the multiple construction defect causal nodes are constructed based on the historical construction defect data. The construction of the causal network for construction deformation is completed by configuring conditional probability allocation based on the frequency of overlap of causal edges between causal nodes.

5. The method for monitoring tunnel surrounding rock deformation as described in claim 4, characterized in that, Based on the surrounding rock deformation early warning, a source tracing analysis is performed to obtain the predicted defects in the construction technology. The method includes: Input the H surrounding rock state associated anomalies in the surrounding rock deformation early warning into the construction deformation causal network, and filter to obtain H construction deformation causal paths; Based on the H causal paths of construction deformation, construction techniques are extracted to obtain H sets of construction techniques and H sets of causal condition probabilities. Interactively obtain H historical construction times for the H groups of construction technologies; Based on the H groups of causal conditional probabilities and H historical construction times, the risk contribution of the H groups of construction technologies is evaluated, and W construction technologies with risk contributions higher than a preset threshold are obtained as predicted defects of the construction technologies.

6. A tunnel surrounding rock deformation monitoring system, characterized in that, The system is used to implement the tunnel surrounding rock deformation monitoring method according to any one of claims 1-5, the system comprising: The surrounding rock state association set acquisition module is used to interactively obtain the surrounding rock state association set, wherein the surrounding rock state association set includes multiple surrounding rock state association indicators. A surrounding rock deformation monitoring array generation module is used to configure monitoring equipment in the tunnel construction area according to the surrounding rock state association set and generate a surrounding rock deformation monitoring array, wherein the surrounding rock deformation monitoring array includes multiple surrounding rock state monitoring devices corresponding to the multiple surrounding rock state association indicators; A communication link construction module is used to pre-build a surrounding rock deformation monitoring cloud and construct a communication link between the surrounding rock deformation monitoring array and the surrounding rock deformation monitoring cloud. The construction technology prediction defect acquisition module is used to perform source analysis based on the surrounding rock deformation warning when the surrounding rock deformation monitoring cloud receives and analyzes the data transmitted back from the surrounding rock deformation monitoring array and outputs a surrounding rock deformation warning; The construction deviation prediction result acquisition module is used to locate construction deviations based on the predicted defects of the construction technology and obtain construction deviation prediction results. The construction early warning instruction generation module is used to generate construction early warning instructions based on the construction deviation prediction results and send the construction early warning instructions to the construction operation and maintenance center.

Images