A digital twin-based metro tunnel dome health monitoring system

Abstract

This application relates to the field of data processing technology, specifically to a digital twin-based subway tunnel dome health monitoring system. The system acquires the spatial topological relationship of the tunnel dome and monitoring data from multiple acquisition cycles. Based on the sensor vibration signal sequences and point cloud data sequences of each monitoring unit across multiple acquisition cycles, it determines the local damage parameters of each monitoring unit. Furthermore, based on the local damage parameters of each monitoring unit and the spatial topological relationship between monitoring units, it determines the damage transmission parameters of the tunnel dome. Finally, based on the local damage parameters of each monitoring unit and the damage transmission parameters of the tunnel dome, it determines the overall health risk index of the tunnel dome. This system can improve the accuracy of subway tunnel dome health monitoring and enhance the risk prediction capability of subway tunnel dome health monitoring.

Description

Technical Field

[0001] This application relates to the field of data processing technology, specifically to a digital twin-based subway tunnel dome health monitoring system. Background Technology

[0002] As the lifeline of urban public transportation, the health of subway tunnels, especially the tunnel dome which bears complex loads for a long time, directly affects operational safety and lifespan. Traditional manual inspection methods are limited by cycle, field of vision, and subjective judgment, making it difficult to achieve comprehensive and real-time structural condition perception and safety early warning. Therefore, the development of intelligent and automated health monitoring systems has become an urgent need to ensure the long-term safe operation of tunnels.

[0003] Currently, the monitoring and surveillance methods for tunnel domes typically involve deploying various sensors, such as vibration and deformation sensors, at key locations within the tunnel dome to collect real-time data on structural physical quantities and then conducting health monitoring based on preset thresholds.

[0004] However, the above-mentioned method of health monitoring by setting a threshold can only passively respond to obvious abnormalities that have occurred. It cannot identify the transmission and diffusion patterns of local damage, and it is difficult to predict the structural risks of the entire area caused by local micro-damage. Therefore, it cannot meet the requirements of refined and intelligent safety operation and maintenance of subway tunnel domes. Summary of the Invention

[0005] To address the technical problem of difficulty in predicting the overall structural risks caused by local micro-damage in existing technologies, this application aims to provide a digital twin-based subway tunnel dome health monitoring system. The specific technical solution adopted is as follows:

[0006] A digital twin-based subway tunnel dome health monitoring system includes a data acquisition unit, a local damage analysis unit, a damage transmission analysis unit, and a health assessment unit. The data acquisition unit acquires the spatial topological relationship of the tunnel dome and monitoring data from multiple acquisition cycles. The spatial topological relationship includes the topological relationship between multiple monitoring units, and the monitoring data from multiple acquisition cycles includes point cloud data sequences and sensor vibration signal sequences for each monitoring unit. The local damage analysis unit determines the local damage parameters of each monitoring unit based on the sensor vibration signal sequences and point cloud data sequences from multiple acquisition cycles. These local damage parameters characterize the structural damage state of a single monitoring unit. The damage transmission analysis unit determines the damage transmission parameters of the tunnel dome based on the local damage parameters of each monitoring unit and the spatial topological relationship between the monitoring units. These damage transmission parameters characterize the diffusion trend of local damage in the tunnel dome structure. The health assessment unit determines the overall health risk index of the tunnel dome based on the local damage parameters of each monitoring unit and the damage transmission parameters of the tunnel dome. This overall health risk index characterizes the risk level of the tunnel dome developing from local damage to overall instability.

[0007] Optionally, the local damage parameter includes damage sensitivity, and the local damage analysis unit includes a damage sensitivity analysis subunit. The damage sensitivity analysis subunit is used to: determine the vibration energy concentration of the first monitoring unit based on the sensor vibration signal sequence of the first monitoring unit in multiple acquisition cycles, wherein the vibration energy concentration is used to characterize the degree of concentration of energy distribution in the frequency domain, and the first monitoring unit is any one of multiple monitoring units; determine the cumulative deformation value of the first monitoring unit based on the point cloud data sequence of the first monitoring unit in multiple acquisition cycles; and determine the damage sensitivity of the first monitoring unit based on the vibration energy concentration and the cumulative deformation value, wherein the damage sensitivity is used to characterize the degree of damage response sensitivity of the monitoring unit to external vibration excitation.

[0008] Optionally, the damage sensitivity analysis subunit is further configured to: determine the deformation sequence based on the coordinate changes of matching points in the point cloud data sequence of adjacent acquisition cycles; perform linear regression analysis on the deformation sequence to obtain the cumulative trend slope; perform first-order differencing on the deformation sequence and calculate the proportion of positive differences and the mean of the absolute values ​​of the differences; determine the deformation irreversibility coefficient based on the proportion of positive differences and the mean of the absolute values ​​of the differences; and determine the cumulative deformation value based on the cumulative trend slope and the deformation irreversibility coefficient.

[0009] Optionally, the local damage parameter also includes a comprehensive damage confidence level. The local damage analysis unit includes a comprehensive damage analysis subunit, which is used to: analyze the correlation between the sensor vibration signal sequence and the point cloud data sequence of the first monitoring unit in multiple acquisition cycles to obtain the causal intensity index of the first monitoring unit. The causal intensity index is used to characterize the strength of the physical causal relationship between the vibration excitation and deformation response of the monitoring unit; and determine the comprehensive damage confidence level of the first monitoring unit based on the damage sensitivity and causal intensity index of the first monitoring unit. The comprehensive damage confidence level is used to characterize the comprehensive credibility of the monitoring unit having structural damage in the current state.

[0010] Optionally, the damage transmission parameter includes a damage transmission intensity factor, and the damage transmission analysis unit includes a transmission intensity analysis subunit, which is used to: determine the neighboring monitoring units of the first monitoring unit based on the spatial topological relationship; determine the structural equivalent stiffness of the first monitoring unit based on the cumulative slope of the deformation sequence of the first monitoring unit; determine the damage risk transmission weight of the first monitoring unit to each neighboring monitoring unit based on the spatial distance between each neighboring monitoring unit and the first monitoring unit and the structural equivalent stiffness of the first monitoring unit; and determine the damage transmission intensity factor of the first monitoring unit based on the comprehensive damage confidence of the first monitoring unit, the damage risk transmission weight of the first monitoring unit to each neighboring monitoring unit, and the damage sensitivity of each neighboring monitoring unit.

[0011] Optionally, the damage transmission parameter further includes an amplification cycle intensity index. The damage transmission analysis unit includes an amplification cycle analysis subunit, which is used to: obtain the damage transmission intensity factor and comprehensive damage confidence of each monitoring unit in multiple acquisition cycles; determine the temporal stability of damage transmission based on the temporal variation trend of the damage transmission intensity factor of each monitoring unit in multiple acquisition cycles; determine the spatial stability of damage transmission based on the spatial variation trend of the damage transmission intensity factor of multiple monitoring units in each acquisition cycle; determine the correlation coefficient of damage transmission amplification effect based on the damage transmission intensity factor and comprehensive damage confidence of each monitoring unit in multiple acquisition cycles; and determine the amplification cycle intensity index based on the damage transmission temporal stability, the damage transmission spatial stability, and the correlation coefficient of damage transmission amplification effect.

[0012] Optionally, the amplified loop analysis subunit is further configured to: determine the set of conduction hotspot monitoring units for each acquisition cycle based on the damage conduction intensity factor for each acquisition cycle; cluster the positions of all conduction hotspot monitoring units in the set of conduction hotspot monitoring units for each acquisition cycle to obtain the cluster center for each acquisition cycle; and determine the spatial stability of damage conduction based on the spatial distance between the cluster centers in adjacent acquisition cycles.

[0013] Optionally, the damage conduction parameter further includes comprehensive damage conduction stability. The damage conduction analysis unit includes a conduction stability analysis subunit, which is used to determine the comprehensive damage conduction stability based on the damage conduction time stability, the damage conduction spatial stability, and the amplification cycle intensity index.

[0014] Optionally, the health assessment unit is also used to: input the global health risk index into a preset health status assessment model to obtain a health risk score for the subway tunnel dome; and issue an early warning signal based on the health risk score.

[0015] Optionally, the data acquisition unit is also used to: divide the surface of the tunnel dome into multiple regular grid areas based on a preset size, and define each grid area as a monitoring unit.

[0016] This application has the following beneficial effects:

[0017] The digital twin-based subway tunnel dome health monitoring system provided in this application analyzes discrete monitoring data within a unified spatial topology. It first analyzes the local damage of each monitoring unit, then analyzes the efficiency of damage propagation in each locality. This allows the assessment results to move beyond isolated monitoring units and reflect the dynamic behavior of damage within the structural spatial network, improving the accuracy of subway tunnel dome health monitoring. The resulting global health risk index enables the perception of microscopic damage to subway tunnel structures and the prediction of macroscopic instability risks, enhancing the risk prediction capability of subway tunnel dome health monitoring and achieving visualized and intelligent monitoring of the entire process from local damage to global instability. Attached Figure Description

[0018] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A structural diagram of a digital twin-based subway tunnel dome health monitoring system provided in one embodiment of this application;

[0020] Figure 2 This is a structural diagram of another digital twin-based subway tunnel dome health monitoring system provided in one embodiment of this application;

[0021] Figure 3 This is a structural diagram of another digital twin-based subway tunnel dome health monitoring system provided in one embodiment of this application. Detailed Implementation

[0022] To further illustrate the technical means and effects adopted by this application to achieve the intended purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a digital twin-based subway tunnel dome health monitoring system proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0023] Unless otherwise defined, 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 application pertains.

[0024] Subway tunnel dome structures bear complex loads over long periods, making comprehensive real-time monitoring difficult using traditional methods. A health monitoring system based on digital twin technology integrates sensor networks and 3D modeling to construct a dynamic virtual mapping of the tunnel structure. This system can fuse multi-dimensional data such as settlement, cracks, and leakage in real time, achieving precise visualization of the dome's condition and providing early warnings of anomalies, significantly improving the level of structural safety maintenance and risk prediction capabilities.

[0025] Current tunnel health monitoring technologies mainly rely on threshold alarms based on statistical analysis and data-driven machine learning assessments. They employ multiple independent sensors deployed in a dispersed manner, with different data formats and sampling frequencies across systems, making data fusion difficult. This results in fragmented monitoring information, making it impossible to construct a comprehensive and coordinated panoramic view of the tunnel dome structure's overall health status. It is also difficult to identify the chain effects caused by local damage, thus hindering more comprehensive health monitoring of the tunnel.

[0026] The following description, in conjunction with the accompanying drawings, details a specific scheme for a digital twin-based subway tunnel dome health monitoring system provided in this application.

[0027] Please see Figure 1 The diagram shows a structural diagram of a digital twin-based subway tunnel dome health monitoring system provided in one embodiment of this application.

[0028] like Figure 1 As shown, the digital twin-based subway tunnel dome health monitoring system 10 includes a data acquisition unit 101, a local damage analysis unit 102, a damage transmission analysis unit 103, and a health assessment unit 104.

[0029] The data acquisition unit 101 is used to acquire the spatial topological relationship of the tunnel dome and monitoring data from multiple acquisition cycles.

[0030] The spatial topology includes the topological relationship between multiple monitoring units, and the monitoring data from multiple acquisition cycles includes the point cloud data sequence and sensor vibration signal sequence of each monitoring unit.

[0031] It should be understood that the data acquisition unit 101 consists of a track-mounted mobile acquisition vehicle, an on-board industrial control computer, a 3D laser scanner, a panoramic camera, sensors, and a data integration unit.

[0032] The track-mounted mobile data acquisition vehicle is equipped with an onboard industrial control computer, a 3D laser scanner, and a panoramic camera. During the tunnel's scheduled maintenance window (i.e., one data acquisition cycle), the track-mounted mobile data acquisition vehicle travels at a constant speed along the track.

[0033] The vehicle-mounted industrial control computer acts as the core controller, triggering and controlling the 3D laser scanner and panoramic camera to collect data synchronously, ensuring that the 3D laser scanner and panoramic camera collect data at the same spatial location and at the same time.

[0034] A 3D laser scanner continuously scans the surface of the tunnel dome and lining structure to obtain high-precision 3D spatial coordinate point cloud data; a panoramic camera works in conjunction with the 3D laser scanner to simultaneously acquire high-resolution panoramic image data of the tunnel surface, which is used to record the appearance and texture information of the structure.

[0035] The sensor is a fiber optic grating sensor network pre-embedded in key stress-bearing parts and vulnerable areas of the tunnel dome. This fiber optic grating sensor network can collect vibration response signals in real time and form the original sensor vibration signal sequence.

[0036] In this embodiment, when the track-mounted mobile data acquisition vehicle travels through the deployment section of the fiber Bragg grating sensor network, the onboard industrial control computer records the precise moment when it passes the spatial position corresponding to each monitoring unit grid. Using this moment as the center, a preset time window (e.g., 0.5 seconds) is extended forward and backward. Vibration signal segments within the preset time window are extracted from the original sensor vibration signal sequence. These extracted vibration signal segments are arranged in chronological order to form the current sensor vibration signal sequence for that monitoring unit in that acquisition cycle. The sensor vibration signals retrieved from the server for multiple acquisition cycles are arranged in chronological order to obtain the sensor vibration signal sequence for that monitoring unit.

[0037] The vehicle-mounted industrial control computer generates a globally unified timestamp for each complete acquisition event, mapping the coordinates of the point cloud data acquired by the 3D laser scanner, the pixel coordinates of the image data acquired by the panoramic camera, and the coordinates of the physical monitoring points corresponding to the sensor vibration signals retrieved from the server to a unified tunnel 3D design model coordinate system, thereby establishing a three-in-one spatial identification code of "geometric coordinates-texture pixels-monitoring points".

[0038] Optionally, the surface of the tunnel dome can be divided into multiple regular grid regions based on a preset size, and each grid region can be defined as a monitoring unit. For one acquisition cycle, the panoramic point cloud data acquired by the 3D laser scanner is mapped to the corresponding monitoring unit grid according to its spatial coordinates. The set of coordinates of all 3D points falling within the same grid is the single-period point cloud data of that monitoring unit in that acquisition cycle. By traversing multiple acquisition cycles in sequence, multiple single-period point cloud data of each monitoring unit can be obtained in chronological order, forming the point cloud data sequence of that monitoring unit.

[0039] It should be understood that all monitoring units and their adjacency relationships together constitute spatial topology.

[0040] For example, a grid area can be 1m × 1m in size, and the multiple acquisition cycles can be 4 acquisition cycles.

[0041] The local damage analysis unit 102 is used to determine the local damage parameters of each monitoring unit based on the sensor vibration signal sequence and point cloud data sequence of each monitoring unit in multiple acquisition cycles.

[0042] The local damage parameter is used to characterize the structural damage state of a single monitoring unit.

[0043] It should be understood that local damage parameters are a type of parameter, namely, parameters used to characterize local damage. Local damage parameters include damage sensitivity and comprehensive damage confidence. The damage sensitivity is used to characterize the sensitivity of the monitoring unit to damage response to external vibration excitation, and the comprehensive damage confidence is used to characterize the overall credibility of the monitoring unit in the current state of structural damage.

[0044] It should be understood that the vibration signal sequence from the sensor records the dynamic response of the structure at its location under external excitations (such as train operation or geological activity). When a structure is damaged, its local stiffness, mass, or damping characteristics change, which directly affects its vibration characteristics, manifesting as changes in the distribution of vibration energy at specific frequencies and alterations in vibration modes. Therefore, by analyzing the vibration signal sequence, abnormal changes in the internal mechanical properties of the structure can be detected.

[0045] Understandably, the accumulation of any structural damage will eventually manifest as irreversible macroscopic or microscopic geometric deformations, such as localized subsidence, convergence, bulging, or subtle changes in surface texture. By analyzing the differences in point cloud data sequences, the magnitude, rate, and trend of deformation can be quantified, thereby capturing the external manifestation of damage in its geometric form.

[0046] The damage transmission analysis unit 103 is used to determine the damage transmission parameters of the tunnel dome based on the local damage parameters of each monitoring unit and the spatial topological relationship between the monitoring units.

[0047] The damage transmission parameter is used to characterize the diffusion trend of local damage in the tunnel dome structure.

[0048] It should be understood that damage transmission parameters are also a type of parameter, namely, parameters used to characterize damage transmission. These damage transmission parameters include the damage transmission intensity factor, the overall stability of damage transmission, and the amplification cycle intensity index. The damage transmission intensity factor is used to characterize the ability and intensity of local damage in a single monitoring unit to spread to neighboring monitoring units; the overall stability of damage transmission is used to characterize the overall stability of the damage transmission process in the tunnel dome structure in terms of temporal evolution and spatial distribution; and the amplification cycle intensity index is used to characterize the strength of the vicious cycle effect formed by the mutual promotion between damage transmission intensity and local damage degree in the tunnel dome structure.

[0049] It is understandable that the stress field inside the tunnel dome is continuous, and any local damage (such as stiffness reduction or crack initiation) will change the stress distribution around it, causing a redistribution of the load path, which may transfer the risk to adjacent areas.

[0050] It should be understood that spatial topology represents the network of physical channels and paths through which damage may propagate.

[0051] Alternatively, the tunnel dome structure can be viewed as a networked system with monitoring units as nodes and spatial topology as connecting edges. The "state" of each node (i.e., its local damage parameters) not only describes its own health but also serves as a potential "disturbance source." The connection relationship between nodes (i.e., spatial topology) determines the potential direction and intensity of this disturbance propagation along the network. By "loading" the parameters reflecting the local state onto the topological network reflecting the spatial structure and conducting systematic spatial correlation analysis, it is possible to simulate and quantify the diffusion trend and impact range of local damage in the structure.

[0052] Health assessment unit 104 is used to determine the overall health risk index of the tunnel dome based on the local damage parameters of each monitoring unit and the damage transmission parameters of the tunnel dome.

[0053] The global health risk index is used to characterize the risk of the tunnel dome developing from local damage to global instability.

[0054] It should be understood that the more severe the local injury and the faster the local damage spreads, the higher the overall health risk index.

[0055] In one optional implementation, the health assessment unit 104 is further used to input the global health risk index into a preset health status assessment model to obtain a health risk score for the subway tunnel dome; and to issue an early warning signal based on the health risk score.

[0056] It should be understood that the higher the overall health risk index and the higher the health risk score, the greater the degree of risk and the more severe the warning signal.

[0057] It should be understood that the preset health status assessment model is a deep learning model (such as convolutional neural networks, temporal Transformers, etc.) used to map multi-dimensional quantitative indicators (such as the global health risk index) into health risk scores.

[0058] Optionally, other traditional monitoring indicators for the current collection period (such as historical settlement data sequences, temporal changes in crack width, density of leakage points, and aging parameters of lining materials) can be obtained and simultaneously input into a preset health status assessment model to obtain the health risk score.

[0059] Optionally, before assessing the health risk score based on the preset health status assessment model, the model needs to be trained. The training process is as follows: based on a large amount of historical case data (i.e., historical global health risk index and traditional monitoring indicators) and the corresponding real health risk scores of tunnels, which have been verified by expert experience or actual engineering projects; through supervised learning algorithms, the model learns the nonlinear mapping relationship between input features and health status levels.

[0060] Optionally, four warning intervals can be divided based on the health risk score, with each warning interval corresponding to a warning signal.

[0061] For example, the health risk score ranges from 0 to 100, and the four warning intervals can be [0, 25), [25, 50), [50, 75), and [75, 100]. The warning signals can be green warning lights, blue warning lights, yellow warning lights, and red warning lights, respectively.

[0062] It should be noted that the different colored warning lights indicate the severity of the warning in descending order, with the red warning light indicating the highest level of risk.

[0063] The digital twin-based subway tunnel dome health monitoring system provided in this application analyzes discrete monitoring data within a unified spatial topology. It first analyzes the local damage of each monitoring unit, then analyzes the efficiency of damage propagation in each locality. This allows the assessment results to no longer be limited to isolated monitoring units, but to reflect the dynamic behavior of damage within the structural spatial network, improving the accuracy of subway tunnel dome health monitoring. The resulting global health risk index enables the perception of microscopic damage to subway tunnel structures and the prediction of macroscopic instability risks, enhancing the risk prediction capability of subway tunnel dome health monitoring and achieving visualized and intelligent monitoring of the entire process from local damage to global instability.

[0064] Subsequently, by synchronously collecting data, various types of fragmented data are extracted into a global health risk index with clear physical meaning. This enables the model to learn more accurately how to assess health risk scores based on the preset health status assessment model, and thus accurately and quickly assess the health status of the tunnel dome.

[0065] Combination Figure 1 ,like Figure 2 As shown, the local damage analysis unit includes a damage sensitivity analysis subunit 1021 and a comprehensive damage analysis subunit 1022.

[0066] The damage sensitivity analysis subunit 1021 is used to perform S201-S203.

[0067] S201. Determine the vibration energy concentration of the first monitoring unit based on the sensor vibration signal sequence of the first monitoring unit in the current acquisition cycle.

[0068] The vibration energy concentration is used to characterize the degree of concentration of energy in the frequency domain, and the first monitoring unit is any one of multiple monitoring units.

[0069] First, power spectral density analysis can be performed on the vibration signal sequence of the sensor to obtain the corresponding power spectrum. Based on the power spectrum, the spectral entropy of the vibration signal sequence of the sensor can be determined. The spectral entropy is a parameter used to characterize the degree of disorder in the frequency domain distribution of vibration energy. The smaller the spectral entropy value, the more concentrated the vibration energy distribution in the frequency domain, the stronger the targeting of external and external vibration excitation in the first monitoring unit area, and the easier it is to induce the risk of local resonance and damage the structure. The higher the entropy value, the more dispersed the energy distribution.

[0070] The vibration energy concentration is calculated based on the maximum spectral entropy when the vibration energy is theoretically uniformly distributed and the spectral entropy of the sensor's vibration signal sequence.

[0071] Optionally, the vibration energy concentration satisfies the following formula:

[0072]

[0073] in, Indicates monitoring unit The concentration of vibrational energy Indicates monitoring unit The spectral entropy corresponding to the sensor vibration signal sequence. This represents the maximum spectral entropy when the vibrational energy is theoretically uniformly distributed, used for normalization. .

[0074] It should be understood that the closer the vibration energy concentration is to 1, the stronger the monitoring unit... The more concentrated the vibrational energy is in the frequency domain, the higher the likelihood of structural resonance caused by external vibration excitation.

[0075] S202. Based on the point cloud data sequence of the first monitoring unit in multiple acquisition cycles, determine the cumulative deformation value of the first monitoring unit.

[0076] It should be understood that the cumulative deformation value is used to characterize the degree of continuous and harmful change in the structural geometry of the first monitoring unit toward damage.

[0077] In one alternative implementation, the deformation sequence can be determined based on the coordinate changes of matching points in the point cloud data sequence of adjacent acquisition cycles; linear regression analysis is performed on the deformation sequence to obtain the cumulative trend slope; first-order differencing is performed on the deformation sequence to calculate the proportion of positive differences and the mean of the absolute values ​​of the differences; the deformation irreversibility coefficient is determined based on the proportion of positive differences and the mean of the absolute values ​​of the differences; and the cumulative deformation value is determined based on the cumulative trend slope and the deformation irreversibility coefficient.

[0078] Optionally, point cloud data sequences from historical adjacent acquisition cycles can be acquired. Based on a preset registration algorithm (such as the iterative nearest point algorithm), spatial location matching and alignment are performed on the point cloud data corresponding to the monitoring unit within every two adjacent acquisition cycles. The Euclidean distance between all successfully matched point pairs within the monitoring unit is calculated, and the deformation of each matched point pair is determined. The average of the deformation sets of all matched point pairs within adjacent acquisition cycles is calculated to obtain the overall deformation of every two adjacent acquisition cycles. The overall deformation of multiple adjacent acquisition cycles is arranged in chronological order to obtain the deformation sequence of the monitoring unit.

[0079] Then, the deformation sequence is fitted using the linear least squares method to obtain the slope of the fitted curve. The slope of the fitted curve is determined as the trend cumulative slope. The trend cumulative slope is used to characterize the average cumulative rate and direction of deformation of the monitoring unit. When the trend cumulative slope is greater than 0, it indicates that the deformation shows an increasing trend with the period, that is, there is continuous deformation accumulation. The larger the absolute value of the trend cumulative slope, the faster the accumulation rate.

[0080] It should be understood that the deformation irreversibility coefficient is used to characterize the average strength of the irreversible plastic deformation portion during the deformation process. The larger the value, the more significant the cumulative effect of irreversible deformation experienced by the monitoring unit.

[0081] It should be understood that the proportion of positive differences is the ratio between the number of positive differences in the first-order differences and the total number of differences, and the mean of the absolute values ​​of the differences is the mean of the absolute values ​​of all first-order differences.

[0082] Optionally, the product of the positive difference percentage and the mean of the absolute values ​​of the differences can be determined as the deformation irreversibility coefficient.

[0083] Optionally, the cumulative trend slope and the deformation irreversibility coefficient can be normalized using the maximum-minimum normalization method, and then the product of the two can be determined as the cumulative deformation value.

[0084] S203. Based on the vibration energy concentration and deformation accumulation, determine the damage sensitivity of the first monitoring unit.

[0085] Optionally, the vibration energy concentration and deformation accumulation can be summed, and the summed value can be normalized again to obtain the damage sensitivity of the first monitoring unit.

[0086] Understandably, the damage sensitivity analysis subunit, by executing the S201-S203 method, correlates and integrates the influence of external excitation (i.e., vibration energy concentration) with the internal deterioration of the structure (i.e., deformation accumulation) to obtain damage sensitivity. This overcomes the limitations of traditional monitoring that only focuses on a single physical quantity (such as only looking at settlement or only analyzing vibration). Starting from the physical essence of "excitation-response" coupling, it identifies weak parts that are prone to damage accumulation under adverse vibration environments earlier and more reliably.

[0087] The comprehensive damage analysis subunit 1022 is used to execute S301-S302.

[0088] S301. Analyze the correlation between the sensor vibration signal sequence and the point cloud data sequence of the first monitoring unit in multiple acquisition cycles to obtain the causal intensity index of the first monitoring unit.

[0089] The causal strength index is used to characterize the strength of the physical causal relationship between the vibration excitation and deformation response of the monitoring unit.

[0090] Optionally, the causal strength index can be calculated by constructing multiple stimulus-response event pairs and analyzing their temporal correlation.

[0091] It should be understood that the current sensor vibration signal sequence of a data acquisition cycle is used to characterize the dynamic excitation (i.e., vibration excitation) applied to the monitoring unit by the track-mounted mobile data acquisition vehicle as it passes by. The deformation of the structure after vibration excitation usually exhibits a hysteresis effect, and multiple data acquisition cycles are required to observe significant cumulative deformation. Furthermore, the cumulative deformation differs for different hysteresis times (i.e., time delays). Therefore, for the excitation event in each data acquisition cycle (e.g., the first data acquisition cycle), the cumulative deformation corresponding to different time delays can be analyzed separately. Then, the most relevant time delay can be determined, and based on the most relevant time delay, the causality strength index can be determined.

[0092] It is understandable that the process of performing feature analysis according to different time delays is essentially the process of calculating the vibration excitation intensity under one acquisition cycle, and then calculating the deformation of one acquisition cycle at different cycle lengths.

[0093] Optionally, the delay parameter can be set to three, such as 1 acquisition cycle, 2 acquisition cycles, and 3 acquisition cycles. For each delay (e.g., 1 acquisition cycle), the root mean square value of the current sensor vibration signal sequence of one acquisition cycle (e.g., the first acquisition cycle) is determined as the vibration excitation intensity, and the deformation between the point cloud data of the next acquisition cycle and the point cloud data of the first acquisition cycle is determined as the current deformation.

[0094] For example, assuming the time delay is one acquisition cycle, and the multiple acquisition cycles are the first acquisition cycle, the second acquisition cycle, the third acquisition cycle, and the fourth acquisition cycle, then the vibration excitation intensity of the first acquisition cycle can be calculated, and the deformation between the point cloud data of the second acquisition cycle and the first acquisition cycle can be calculated.

[0095] By iterating through each acquisition cycle, the vibration intensity sequence and the current deformation sequence are obtained. Based on the Pearson correlation coefficient formula, the cross-correlation coefficient between the vibration intensity sequence and the current deformation sequence at each time delay is calculated. This cross-correlation coefficient is used to quantify the degree of linear correlation between the two signals under each time delay. A cross-correlation coefficient sequence is obtained by iterating through all time delays.

[0096] Next, the maximum value in the cross-correlation coefficient sequence is identified and denoted as the maximum cross-correlation coefficient. Simultaneously, the time delay corresponding to this maximum cross-correlation coefficient is recorded and denoted as the peak delay. This peak delay is then normalized to obtain the normalized peak delay. The peak delay characterizes the time interval between the strongest correlation between the vibration excitation and the deformation response.

[0097] It should be noted that due to this lag effect, there may be some acquisition cycles that cannot obtain the current deformation variables corresponding to the time delay. For example, the vibration excitation intensity of the last acquisition cycle cannot obtain the current deformation variables one acquisition cycle later. When traversing each acquisition cycle, the acquisition cycle that cannot obtain the current deformation variables corresponding to the time delay is not included in the statistics.

[0098] Optionally, a normalization constant in the time dimension, such as 1 second, can be set, and the ratio of the peak delay to the normalization constant can be determined as the normalized peak delay.

[0099] Optionally, the causality strength index satisfies the following formula:

[0100]

[0101] in, Indicates monitoring unit The causal strength index Indicates monitoring unit The maximum cross-correlation coefficient, Indicates monitoring unit Peak latency after normalization Represents an exponential function. The larger, The closer it is to 0.

[0102] It should be understood that The value range is [0,1]. The closer the value is to 1, the stronger the correlation and the shorter the response lag time, which conforms to the physical causal chain of "excitation-response". The closer the value is to 0, the weaker the correlation or the earlier the response, and the physical causal relationship does not hold.

[0103] S302. Based on the damage sensitivity and causal intensity index of the first monitoring unit, determine the comprehensive damage confidence level of the first monitoring unit.

[0104] Optionally, the overall damage confidence level satisfies the following formula:

[0105]

[0106] in, Indicates monitoring unit The overall damage confidence level, Indicates monitoring unit The damage sensitivity, Indicates monitoring unit The causal strength index.

[0107] In this formula, This can be understood as a gain term with a value range of [1, 2]. Damage sensitivity defines the baseline of damage probability, while the causality intensity index corrects or amplifies the credibility of this baseline probability. When the causality intensity index is very high (close to 1), it indicates that the observed deformation is likely directly caused by vibration excitation, greatly enhancing the credibility of the existence of real structural damage at that location. Therefore, through... The gain significantly improves the final confidence level; conversely, if the causal strength index is very low (close to 0), even if the damage sensitivity is high, it may just be caused by other non-structural factors or noise, so the gain is very small (close to 1), and the final confidence level depends mainly on the damage sensitivity itself.

[0108] It should be understood that the overall damage confidence calculated above is used to represent the overall damage confidence at the current moment (or the current acquisition period).

[0109] The comprehensive damage analysis subunit 1022, by executing S301-S302, analyzes the temporal correlation between vibration signal and deformation response (especially the characteristic that deformation lags behind vibration), so that the determined "comprehensive damage confidence" not only reflects the possibility of damage (sensitivity), but also enhances the credibility of the judgment (causal strength).

[0110] Combination Figure 2 ,like Figure 3 As shown, the damage conduction analysis unit 103 includes a conduction intensity analysis subunit 1031, an amplification cycle analysis subunit 1032, and a conduction stability analysis subunit 1033.

[0111] The conduction strength analysis subunit 1031 is used to perform S401-S404.

[0112] S401. Based on spatial topological relationships, determine the neighboring monitoring units of the first monitoring unit.

[0113] Optionally, based on the 3×3 neighborhood rule, the eight neighboring monitoring units directly above, below, to the left, to the right, and along the four diagonal directions of the first monitoring unit can be determined as neighborhood monitoring units.

[0114] Optionally, if the first monitoring unit is located at the edge of the tunnel dome grid, its neighborhood monitoring unit set only includes the actual adjacent units.

[0115] S402. Based on the cumulative slope of the deformation sequence of the first monitoring unit, determine the structural equivalent stiffness of the first monitoring unit.

[0116] It should be understood that the equivalent stiffness of this structure is used to characterize the ability of the local structure where the monitoring unit is located to resist deformation.

[0117] It is understandable that the cumulative slope of the trend obtained above represents the average cumulative rate of deformation of the monitoring unit. The larger the cumulative slope of the trend, the faster the deformation of the monitoring unit develops and the weaker its ability to resist deformation, that is, the smaller the equivalent stiffness of the structure.

[0118] Optionally, the equivalent stiffness of the structure satisfies the following formula:

[0119]

[0120] in, Indicates monitoring unit The structural equivalent stiffness, Indicates monitoring unit The cumulative slope of the trend of the deformation sequence, Represents a very small positive number, such as 10. -6 Used to prevent division by zero in the denominator. This indicates taking the absolute value.

[0121] S403. Based on the spatial distance between each neighboring monitoring unit and the first monitoring unit and the structural equivalent stiffness of the first monitoring unit, determine the damage risk transmission weight of the first monitoring unit to each neighboring monitoring unit.

[0122] Optionally, the Euclidean distance between the three-dimensional coordinates of the center points of the two monitoring units can be used to determine the spatial distance between the two monitoring units.

[0123] It should be understood that the damage risk transmission weight is used to quantify the relative ease with which damage is transmitted from the source monitoring unit to a specific neighboring unit. The larger the damage risk transmission weight, the better the "accessibility" of the transmission path, and the easier it is for damage to spread in this direction.

[0124] Optionally, the damage risk transmission weights satisfy the following formula:

[0125]

[0126] in, Indicates monitoring unit Neighborhood monitoring unit Damage risk transmission weight, Indicates monitoring unit With neighboring monitoring units Spatial distance between them Indicates monitoring unit The structural equivalent stiffness.

[0127] S404. Based on the comprehensive damage confidence of the first monitoring unit, the damage risk transmission weight of the first monitoring unit to each neighboring monitoring unit, and the damage sensitivity of each neighboring monitoring unit, determine the damage transmission intensity factor of the first monitoring unit.

[0128] Optionally, the damage conduction intensity factor satisfies the following formula:

[0129]

[0130] in, Indicates monitoring unit Damage conduction intensity factor Indicates monitoring unit The overall damage confidence level, Indicates monitoring unit Neighborhood monitoring unit Damage risk transmission weight, Indicates neighborhood monitoring unit The damage sensitivity, This indicates the number of neighborhood monitoring units.

[0131] In this formula, Representative monitoring unit The amount of damage that exists within itself and is yet to spread. Indicates the neighboring monitoring units to which damage has spread. The level of difficulty Indicates damage from monitoring unit The combined effect of "channel capacity" and "target vulnerability" spreading outwards results in severe damage to the target itself. However, if the surrounding units are far apart, the target has high stiffness (high path resistance), or the surrounding units are very robust (the target is not easily damaged), then the total conduction potential is small, and the overall conduction strength will be limited.

[0132] The conduction strength analysis sub-unit 1031, by executing S401-S404 and combining the equivalent stiffness of the structure, spatial distance and damage sensitivity of neighboring units, accurately characterizes the possibility and intensity of local damage spreading to the surrounding area, solving the problem that traditional monitoring cannot predict the trend of damage spread.

[0133] The amplified loop analysis subunit 1032 is used to execute S501-S505.

[0134] S501. Obtain the damage transmission intensity factor and comprehensive damage confidence of each monitoring unit in multiple acquisition cycles.

[0135] It should be understood that the method for determining the damage conduction intensity factor (or overall damage confidence) of each monitoring unit in other acquisition cycles is the same or similar to the method for determining the damage conduction intensity factor (or overall damage confidence) in the current acquisition cycle, and will not be elaborated here.

[0136] S502. Based on the time variation trend of the damage conduction intensity factor of each monitoring unit in multiple acquisition cycles, determine the time stability of damage conduction.

[0137] First, linear regression analysis is performed on the time series of damage conduction intensity factors for each monitoring unit across multiple acquisition cycles. The slope of the fitted curve for this time series is obtained by fitting it using a preset linear fitting algorithm (such as linear least squares). This slope is determined as the conduction intensity change slope, which characterizes the rate of change of the damage conduction intensity factor of a single monitoring unit with the acquisition cycle. Then, the conduction intensity change slopes of all monitoring units are normalized to obtain the conduction trajectory change rate of each monitoring unit. The closer this conduction trajectory change rate is to 1, the faster the damage conduction intensity factor of the monitoring unit increases over time.

[0138] Then, the arithmetic mean of the rate of change of the conduction trajectory of all monitoring units is calculated, and this mean is determined as the damage conduction time stability.

[0139] S503. Based on the spatial variation trend of damage conduction intensity factor of multiple monitoring units within each acquisition cycle, determine the spatial stability of damage conduction.

[0140] It should be understood that damage conduction spatial stability is used to characterize the stability of the spatial distribution of damage conduction hotspots in the tunnel dome in the recent period (i.e., within multiple acquisition cycles), reflecting whether the conduction hotspots are concentrated in a fixed spatial area, and is a global indicator.

[0141] In one alternative implementation, the set of conduction hotspot monitoring units for each acquisition cycle can be determined based on the damage conduction intensity factor for each acquisition cycle; the positions of all conduction hotspot monitoring units in the set of conduction hotspot monitoring units for each acquisition cycle are clustered to obtain the cluster center for each acquisition cycle; and the spatial stability of the damage conduction is determined based on the spatial distance between the cluster centers in adjacent acquisition cycles.

[0142] Optionally, for any acquisition period (e.g., the first acquisition period), the damage conduction intensity factors of all monitoring units can be sorted in descending order, and the monitoring units ranked in the top preset percentage (e.g., 20%) can be selected as conduction hotspot monitoring units.

[0143] It should be understood that this transmission hotspot monitoring unit is the monitoring unit with the strongest damage spread capability and the highest risk.

[0144] Optionally, clustering can be performed on the three-dimensional spatial coordinates of the center points of all conductive hotspot monitoring units in a set of conductive hotspot monitoring units based on a preset spatial clustering algorithm.

[0145] For example, clustering can be performed based on K-means clustering. The elbow rule is used to analyze the sum of squared errors within clusters under different numbers of clusters to determine the optimal number of clusters K. Then, K-means clustering is performed with K as a parameter to obtain multiple clusters. The centroid of the cluster containing the most coordinate points is used as the cluster center to characterize the main concentrated area of ​​the periodic transmission hotspot.

[0146] It should be understood that the cluster center of a collection cycle reflects the main clustering location of high-risk damage transmission areas in the tunnel dome space within that cycle. By comparing the cluster centers of different cycles, the spatial drift of hotspot areas can be analyzed.

[0147] It should be understood that the greater the spatial distance between cluster centers in adjacent acquisition cycles, the smaller the spatial stability of damage propagation should be.

[0148] Alternatively, the Euclidean distance between cluster centers can be defined as the spatial distance.

[0149] Optionally, the spatial stability of damage conduction satisfies the following formula:

[0150]

[0151] in, This indicates the spatial stability of damage transmission in the tunnel dome. Indicates the first Cluster centers in the first collection cycle and the first Spatial distance of cluster centers over a collection period Indicates the number of multiple acquisition cycles. This represents a normalization function, such as max-min normalization, used to map values ​​to the interval [0, 1].

[0152] In this formula, This represents the mean distance; the larger the mean distance, the smaller the spatial stability of damage conduction.

[0153] S504. Based on the damage conduction intensity factor and comprehensive damage confidence of each monitoring unit in multiple acquisition cycles, determine the correlation coefficient of damage conduction amplification effect.

[0154] First, the damage conduction intensity factor and the overall damage confidence level of the conduction hotspot monitoring unit for each acquisition cycle are averaged to obtain the mean damage conduction intensity factor and the mean overall damage confidence level for each acquisition cycle. The mean damage conduction intensity factor and the mean overall damage confidence level for multiple acquisition cycles are then arranged in chronological order to obtain the mean damage conduction intensity factor sequence and the mean overall damage confidence level sequence.

[0155] Then, the correlation between the two sequences is determined based on a preset correlation analysis algorithm (such as Pearson correlation coefficient), and denoted as the damage conduction amplification effect correlation coefficient.

[0156] It should be understood that when temporal stability is poor (damage conduction intensity factor continues to increase) and spatial stability is poor (conduction path is concentrated), a loop of amplification effect will be formed, which is "damage conduction intensity factor increases → damage intensifies → damage conduction intensity factor further increases", accelerating the evolution of the global structure from "local damage" to "global instability". Therefore, the correlation coefficient of this damage conduction amplification effect is used to amplify this effect.

[0157] S505. Based on the damage conduction time stability, damage conduction spatial stability, and the correlation coefficient of damage conduction amplification effect, determine the amplification cycle intensity index.

[0158] Optionally, the product of the damage conduction time stability, the damage conduction spatial stability, and the correlation coefficient of the damage conduction amplification effect can be normalized and then determined as the amplification cycle intensity index.

[0159] Alternatively, normalization can be performed based on the min-max normalization method.

[0160] The amplified cycle analysis subunit 1032 performs time stability analysis by executing S501-S504, which can capture the dynamic trend of damage transmission intensity and determine whether the risk continues to increase. Spatial stability analysis can reflect the distribution pattern of transmission hotspots and determine whether a fixed risk transmission channel has been formed. By analyzing the temporal correlation between damage transmission intensity and comprehensive damage confidence, the vicious cycle effect of "damage-transmission" is quantified, so that the global risk assessment can not only grasp the current risk status, but also predict the speed of risk evolution and provide early warning of global instability risk caused by vicious cycles.

[0161] The conduction stability analysis subunit 1033 is used to determine the overall stability of damage conduction based on the damage conduction time stability, damage conduction space stability, and amplified cycle intensity index.

[0162] Optionally, the overall stability of damage conduction satisfies the following formula:

[0163]

[0164] in, This indicates the overall stability of damage transmission in the tunnel dome. This indicates the spatial stability of damage transmission in the tunnel dome. This indicates the spatial stability of damage transmission in the tunnel dome. Indicates the amplification cycle strength index. This represents a normalization function, such as max-min normalization, used to map values ​​to the interval [0, 1].

[0165] It should be understood that the range of values ​​for is [0, 1], and after adding 1, it is equal to . and Multiplication can amplify the effect of this cycle.

[0166] The comprehensive stability of damage transmission obtained by this formula integrates temporal stability, spatial stability, and the cyclical promotion relationship between the two. It can more comprehensively grasp the stability and evolution direction of damage diffusion, provide richer evidence for assessing the development trend of global risks, and further improve the comprehensiveness and accuracy of risk warning.

[0167] In one implementation of this application, the health assessment unit 104 can be specifically used to determine the average comprehensive damage confidence level and the average damage transmission intensity factor of multiple monitoring units in the current acquisition cycle, and then determine the global health risk index of the tunnel dome based on the average comprehensive damage confidence level, the average damage transmission intensity factor, the comprehensive stability of damage transmission, and the amplified cycle intensity index.

[0168] It should be understood that the mean of the comprehensive damage confidence level is the average of the comprehensive damage confidence levels of multiple monitoring units in the current acquisition period, and the mean of the damage conduction intensity factor is the average of the damage conduction intensity factors of multiple monitoring units in the current acquisition period.

[0169] Optionally, the overall health risk index satisfies the following formula:

[0170]

[0171] in, Indicates the overall health risk index. This represents the mean confidence level of the overall damage. Indicates the damage conduction intensity factor. Indicates the overall stability of damage transmission. Indicates the amplification cycle strength index. This represents a normalization function, such as max-min normalization, used to map values ​​to the interval [0, 1].

[0172] In this formula, This serves as the risk basis, representing the static foundation or initial stock of risk across the entire domain. This is represented by the risk diffusion amplification term, signifying the dynamic potential and instability of risk spreading from a local to a global scale. This represents the risk cycle accelerator, indicating whether there is a positive feedback mechanism within the structure that leads to self-reinforcing and accelerated escalation of risk. The multiplication of these three factors emphasizes the non-linear nature of risk: extremely high risk only arises when basic damage, diffusion threats, and vicious cycles coexist and are coupled with each other; reducing the risk in any one link will inhibit the risk effects of other links.

[0173] In summary, the digital twin-based subway tunnel dome health monitoring system 10 provided in this application uses a mobile acquisition platform to simultaneously acquire 3D point cloud, panoramic images, and time-series data from physical sensors. By fusing these data using a unified spatiotemporal reference, it effectively overcomes the fragmentation problems caused by scattered and inconsistent data formats in traditional monitoring, constructing a full-domain digital mapping of the tunnel dome structure. Based on this, through the collaborative analysis of vibration signals and deformation sequences, it quantifies the concentration of vibration energy and the trend of deformation accumulation. Combined with a time-series causal verification mechanism, it achieves accurate identification and confidence assessment of local damage, significantly improving the reliability of damage detection and early warning. Furthermore, a transmission network model based on spatial topology and structural stiffness is introduced to quantify the damage transmission intensity factor and path stability between neighboring units, enabling dynamic tracking of damage diffusion trends and identification of potential risk transmission hotspots. By integrating multi-dimensional features such as local damage intensity, transmission amplification effect, and evolution stability, the system achieves a comprehensive quantitative assessment of the overall health status of the tunnel dome and triggers graded early warnings based on dynamic risk levels, providing data support for operation and maintenance decisions. This comprehensively improves the safety operation and maintenance level and risk prediction capability of subway tunnel structures, and realizes visualized and intelligent monitoring of the entire process from local damage to overall instability.

[0174] 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. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0175] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A digital twin-based metro tunnel dome health monitoring system, characterized in that, It includes a data acquisition unit, a local damage analysis unit, a damage transmission analysis unit, and a health assessment unit: The data acquisition unit is used to acquire the spatial topological relationship of the tunnel dome and the monitoring data of multiple acquisition cycles. The spatial topological relationship includes the topological relationship between multiple monitoring units, and the monitoring data of multiple acquisition cycles includes the point cloud data sequence and sensor vibration signal sequence of each monitoring unit. The local damage analysis unit is used to determine the local damage parameters of each monitoring unit based on the sensor vibration signal sequence and point cloud data sequence of each monitoring unit in multiple acquisition cycles. The local damage parameters are used to characterize the structural damage state of a single monitoring unit. The local damage parameters include damage sensitivity and comprehensive damage confidence. The damage sensitivity is used to characterize the sensitivity of the monitoring unit to damage response to external vibration excitation, and the comprehensive damage confidence is used to characterize the comprehensive confidence that the monitoring unit has structural damage in the current state. The damage transmission analysis unit is used to determine the damage transmission parameters of the tunnel dome based on the local damage parameters of each monitoring unit and the spatial topological relationship between the monitoring units. The damage transmission parameters are used to characterize the diffusion trend of local damage in the tunnel dome structure. The damage transmission parameters include damage transmission intensity factor, amplification cycle intensity index, and damage transmission comprehensive stability. The damage transmission intensity factor is used to characterize the ability and intensity of local damage of a single monitoring unit to spread to neighboring monitoring units. The damage transmission comprehensive stability is used to characterize the overall stability of the damage transmission process in the tunnel dome structure in terms of temporal evolution and spatial distribution. The amplification cycle intensity index is used to characterize the strength of the vicious cycle effect formed by the mutual promotion of damage transmission intensity and local damage degree in the tunnel dome structure. A health assessment unit is used to determine the global health risk index of the tunnel dome based on the local damage parameters of each monitoring unit and the damage transmission parameters of the tunnel dome. The global health risk index is used to characterize the degree of risk of the tunnel dome developing from local damage to global instability.

2. The digital-twin-based metro tunnel dome health monitoring system according to claim 1, wherein, The local damage analysis unit includes a damage sensitivity analysis subunit, which is used for: The vibration energy concentration of the first monitoring unit is determined based on the sensor vibration signal sequence of the first monitoring unit in multiple acquisition cycles. The vibration energy concentration is used to characterize the degree of concentration of energy in the frequency domain distribution. The first monitoring unit is any one of multiple monitoring units. Based on the point cloud data sequence of the first monitoring unit in multiple acquisition cycles, the cumulative deformation value of the first monitoring unit is determined; The damage sensitivity of the first monitoring unit is determined based on the vibration energy concentration and the cumulative deformation value.

3. The digital-twin-based metro tunnel dome health monitoring system according to claim 2, wherein, The damage sensitivity analysis subunit is also used for: Deformation sequence is determined based on the coordinate changes of matching points in point cloud data sequences from adjacent acquisition cycles; Linear regression analysis was performed on the deformation sequence to obtain the cumulative trend slope; Perform first-order differencing on the deformation sequence and calculate the proportion of positive differences and the mean of the absolute values ​​of the differences; The deformation irreversibility coefficient is determined based on the positive difference ratio and the mean of the absolute value of the difference. The cumulative deformation value is determined based on the cumulative slope of the trend and the coefficient of irreversible deformation.

4. The digital-twin-based metro tunnel dome health monitoring system according to claim 2, wherein, The local damage analysis unit includes a comprehensive damage analysis subunit, which is used for: The correlation between the sensor vibration signal sequence and the point cloud data sequence of the first monitoring unit in multiple acquisition cycles is analyzed to obtain the causal intensity index of the first monitoring unit. The causal intensity index is used to characterize the physical causal correlation strength between the vibration excitation and deformation response of the monitoring unit. Based on the damage sensitivity and causal strength index of the first monitoring unit, the overall damage confidence level of the first monitoring unit is determined.

5. The digital-twin-based metro tunnel dome health monitoring system according to claim 4, wherein, The damage conduction analysis unit includes a conduction intensity analysis subunit, which is used for: Based on the spatial topological relationship, the neighboring monitoring units of the first monitoring unit are determined; The structural equivalent stiffness of the first monitoring unit is determined based on the cumulative slope of the deformation sequence of the first monitoring unit. Based on the spatial distance between each neighboring monitoring unit and the first monitoring unit, as well as the structural equivalent stiffness of the first monitoring unit, the damage risk transmission weight of the first monitoring unit to each neighboring monitoring unit is determined. The damage transmission intensity factor of the first monitoring unit is determined based on the comprehensive damage confidence of the first monitoring unit, the damage risk transmission weight of the first monitoring unit to each neighboring monitoring unit, and the damage sensitivity of each neighboring monitoring unit.

6. The digital-twin-based metro tunnel dome health monitoring system according to claim 5, wherein, The damage conduction analysis unit includes an amplification loop analysis subunit, which is used for: Obtain the damage conduction intensity factor and overall damage confidence level for each monitoring unit across multiple acquisition cycles; Based on the temporal variation trend of the damage conduction intensity factor of each monitoring unit in multiple acquisition cycles, the damage conduction time stability is determined. Based on the spatial variation trend of damage conduction intensity factor of multiple monitoring units within each acquisition cycle, the spatial stability of damage conduction is determined. Based on the damage conduction intensity factor and comprehensive damage confidence of each monitoring unit in multiple acquisition cycles, the correlation coefficient of damage conduction amplification effect is determined; Based on the damage conduction time stability, the damage conduction spatial stability, and the correlation coefficient of the damage conduction amplification effect, the amplification cycle intensity index is determined.

7. The digital-twin-based metro tunnel dome health monitoring system according to claim 6, wherein, The amplified loop analysis subunit is also used for: Based on the damage conduction intensity factor of each acquisition cycle, the set of conduction hotspot monitoring units for each acquisition cycle is determined; Cluster the locations of all conducted hotspot monitoring units in the set of conducted hotspot monitoring units for each acquisition cycle to obtain the cluster center for each acquisition cycle; The spatial stability of damage transmission is determined based on the spatial distance between cluster centers in adjacent acquisition cycles.

8. The digital twin-based subway tunnel dome health monitoring system according to claim 6, characterized in that, The damage conduction analysis unit includes a conduction stability analysis subunit, which is used for: The overall stability of damage conduction is determined based on the damage conduction time stability, the damage conduction spatial stability, and the amplification cycle intensity index.

9. The digital twin-based subway tunnel dome health monitoring system according to any one of claims 1 to 8, characterized in that, The health assessment unit is also used for: The global health risk index is input into a preset health status assessment model to obtain the health risk score of the subway tunnel dome. A warning signal is issued based on the health risk score.

10. The digital-twin-based metro tunnel dome health monitoring system according to claim 1, wherein, The data acquisition unit is also used for: The surface of the tunnel dome is divided into multiple regular grid areas based on a preset size, and each grid area is defined as a monitoring unit.

Images