Tunnel construction safety intelligent management and control method and system based on dynamic sensing data
By constructing a tunnel perception performance model and dynamic anomaly analysis, combined with finite element and Monte Carlo simulations, the problem of insufficient risk quantification in safety monitoring during tunnel construction was solved, and efficient safety management and control of tunnel construction was achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-15
- Publication Date
- 2026-03-13
AI Technical Summary
Existing tunnel construction safety monitoring methods lack efficient anomaly analysis and risk quantification mechanisms, resulting in inaccurate collapse probability assessments and difficulty in identifying safety hazards in advance, which often leads to accidents, especially under complex geological conditions.
By constructing a tunnel perception model, using vector lines and pressure spheres to represent surrounding rock pressure, crown settlement, and surrounding convergence parameters, dynamic adjustments and anomaly analysis are performed. Combined with finite element analysis and Monte Carlo simulation, anomaly risks are quantified, and real-time data comparison is conducted to achieve intelligent management and control.
It has improved the efficiency and accuracy of safety management and control during tunnel construction, reduced human intervention, and enhanced the ability to identify and warn of potential risks.
Smart Images
Figure CN121659664A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of tunnel management technology, and in particular to a method and system for intelligent management and control of tunnel construction safety based on dynamic sensing data. Background Technology
[0002] In the field of tunnel engineering construction, safety monitoring is a crucial link in ensuring the stability of the construction process and the long-term reliability of the structure. Current technologies primarily rely on sensors to collect parameters such as surrounding rock pressure, crown settlement, and perimeter convergence, and then monitor and issue early warnings through basic information systems. However, these methods have significant shortcomings, particularly the lack of efficient anomaly analysis and risk quantification mechanisms. Existing static models or threshold judgments struggle to handle multi-dimensional time-series data, leading to inaccurate collapse probability assessments and difficulty in identifying safety hazards in advance. This problem is particularly prominent under complex geological conditions, often triggering accidents such as surrounding rock instability and structural deformation, affecting project progress and personnel safety. Summary of the Invention
[0003] The purpose of this invention is to provide a tunnel safety management method and system capable of providing effective early warning.
[0004] This invention discloses an intelligent management and control method for tunnel construction safety based on dynamic sensing data, including: Step S100: Analyze historical tunnel data, including obtaining surrounding rock pressure parameters, crown settlement parameters, and peripheral convergence parameters at different locations on the tunnel face at preset intervals as the tunnel advances. The crown settlement parameters include the unit settlement amount and cumulative settlement value per unit time, and the peripheral convergence parameters include the unit convergence amount and cumulative convergence amount per unit time. Step S200: Lock the tunnel face one by one according to the preset interval, determine the center point of each tunnel face, connect them with each other to form a tunnel trend line, and construct a tunnel face representation component at the corresponding position node of each tunnel face. The tunnel face representation component is used to represent the surrounding rock pressure parameters, arch settlement parameters and surrounding convergence parameters of the tunnel face. The combination of the tunnel trend line and several tunnel face representation components is identified as the tunnel perception representation model. Step S300: Based on various parameters, adjust the tunnel perception performance model to obtain dynamic performance state, analyze several dynamic performance states, and mark the physical location segments and frame number segments with anomalies, and identify the marked dynamic performance state as the reference dynamic performance state. Step S400: Integrate several reference dynamic performance states to obtain a reference dynamic performance state analysis model, and use the reference dynamic performance state analysis model to analyze real-time dynamic sensing data.
[0005] In some embodiments disclosed in this invention, the method for constructing a face representation model includes: Step S201: Determine the detection points set on the edge of the working face, and construct several representation vector lines based on the direction of each detection point relative to the center point. One end of the representation vector line coincides with the center point, and the other end points to the detection point. A pressure representation sphere is set at the end pointing to the detection point. The diameter of the pressure representation sphere is determined according to the surrounding rock pressure at the detection point, and the length of the representation vector line is determined according to the settlement or convergence.
[0006] In some embodiments disclosed in this invention, the method for analyzing dynamic behavior includes: Step S301 involves analyzing each dynamic performance frame in the dynamic performance state, including analyzing the length of the performance vector line in each dynamic performance frame and the diameter of each pressure performance sphere. The analysis method includes: Step S3011: Mark the vector lines whose length changes and whose change amplitude per unit time is greater than or equal to a preset value; Step S3012: Combine several representation vector lines with similar positional relationships, including representation vector lines on the same face representation component and adjacent long face representation components, to form a marked representation vector line group. Step S3013: Determine the ends of the representation vector lines in the marked representation vector line group, and perform surface fitting with the ends as boundaries to obtain the representation vector surface; Step S3014: Save the performance vector surfaces corresponding to several consecutive frames of the dynamic performance state, as well as the corresponding pressure performance spheres, to obtain the performance state marked performance segments.
[0007] In some embodiments disclosed in this invention, the method for identifying abnormal physical segments and frame number segments includes: Step S302: Equivalent expansion of the performance marker segment is performed to construct several performance extended segments. The combination of the performance marker segment and the performance extended segment is denoted as the performance comprehensive segment set. Step S303: Perform anomaly risk analysis on each performance segment in the comprehensive performance segment set, including simulating the surrounding rock pressure parameters, crown settlement parameters and surrounding convergence parameters corresponding to each performance segment, outputting the corresponding collapse probability, and determining the corresponding anomaly risk index based on the preset probability interval to which the collapse probability belongs. Step S304: Perform retrieval feature analysis on each performance segment in the performance comprehensive performance segment set, and classify the performance segments step by step based on the determined retrieval features, and determine the retrieval feature group corresponding to each category group to obtain the performance segment retrieval model.
[0008] In some embodiments disclosed in this invention, the method for simulating the surrounding rock pressure parameters, crown settlement parameters, and surrounding convergence parameters corresponding to each manifested segment includes: Step S3031: Construct a simulation environment, including extracting the corresponding surrounding rock pressure parameter sequence, crown settlement parameter sequence, and perimeter convergence parameter sequence based on the physical location and frame range of the performance segment. The surrounding rock pressure parameter sequence includes the time series of pressure values at each detection point, the crown settlement parameter sequence includes the settlement change sequence, and the perimeter convergence parameter sequence includes the convergence change sequence. Step S3032: Preprocess the extracted parameter sequence, including filtering outliers in the sequence and smoothing it based on a preset time window to obtain a smoothed parameter sequence group. Step S3033: Construct a tunnel collapse simulation model. The model includes a finite element analysis module and a probability statistics module. The finite element analysis module is used to simulate the mechanical effects of surrounding rock pressure, crown settlement and perimeter convergence on the tunnel structure. The probability statistics module is used to calculate conditional probabilities based on historical collapse data. Step S3034: Input the smoothing parameter sequence group into the finite element analysis module to perform mechanical simulation, including calculating the stress field under the surrounding rock pressure distribution, the displacement field caused by the crown settlement and the deformation field caused by the surrounding convergence, and output the structural stability index under the simulation. Step S3035: Input the structural stability index into the probability statistics module to calculate the collapse probability. This includes using the Monte Carlo simulation method to perform multiple iterative simulations of the random disturbance of the parameter sequence and to count the proportion of collapse events as the corresponding collapse probability. Step S3036: Output the collapse probability and map it to the abnormal risk index according to the preset probability interval. The preset probability interval includes a low-risk interval, a medium-risk interval, and a high-risk interval, which correspond to different abnormal risk index values.
[0009] In some embodiments disclosed in this invention, a method for performing feature analysis on each performance segment in the performance comprehensive performance segment set includes: Step S3037: For the vertical direction of the tunnel, construct several levels of surface intercepting planes, analyze the surface intersection edges of each representation vector surface and surface intercepting plane, determine the length of the surface intersection edges and record it as the intersection edge length, record the intersection edge length corresponding to each level of surface intercepting plane, and obtain the intersection edge length sequence. Step S3038: Take each cross edge length in the cross edge length sequence as a retrieval feature that is separated step by step, calculate the average diameter of the pressure representation sphere corresponding to the representation vector surface, and take the average diameter as the retrieval feature.
[0010] In some embodiments disclosed in this invention, the method for analyzing real-time dynamic sensing data using a reference dynamic performance analysis model includes: Step S401: Based on the real-time dynamic sensing data of the tunnel, construct the real-time performance segment, and perform feature analysis on the real-time performance segment to obtain the real-time cross-edge length sequence and the real-time average diameter of the pressure performance sphere. Step S402: Based on the real-time cross-edge length sequence and the real-time average diameter, a search is performed to find several performance segments in the performance comprehensive performance segment set, which are recorded as the performance segments for comparison. Step S403: Compare each comparison performance segment with the real-time performance segment to determine the degree of equivalence between the two, and based on the degree of equivalence, select a reference performance segment from the comparison performance segments.
[0011] In some embodiments disclosed in this invention, the method for calculating the degree of equivalence includes: Step S4031: Perform several mutual attitude adjustments on the representation vector surfaces between the representation segments, including rotating and translating each pair of representation vector surfaces to minimize the geometric differences between the surfaces, and calculate the ratio of the overlapping area of the adjusted surfaces as the attitude adjustment similarity, and select the optimal attitude adjustment method. Step S4032: Extract feature points from the surface of the performance vector corresponding to the optimal posture adjustment method, including identifying key bending points, peak points and valley points on the surface, constructing a feature point matrix, matching and comparing the feature point matrix, calculating the distance between points, and obtaining the corresponding distance matrix between points. Step S4033: Analyze the pressure-representing spheres corresponding to the surface of the representation vector and determine the diameter difference between them to obtain the diameter difference matrix; Step S4034: Perform a comprehensive analysis of the point distance matrix and diameter difference matrix of each frame in the performance segment to determine the degree of equivalence.
[0012] In some embodiments disclosed in this invention, the expression for calculating the degree of equivalence is as follows: ; Where D represents the same degree. The distance matrix between points has an equal influence on the weights. The diameter difference matrix has equal influence weights. To preset the maximum distance between points as an equivalent parameter, To represent the number of matching points in the distance matrix corresponding to the i-th frame of the representation segment, the matching point is determined by the distance between points being less than or equal to a preset value, where h is the total number of points in the distance matrix. The number of matching spheres in the diameter difference matrix corresponding to the i-th frame of the performance segment is represented by q, and the method for judging matching spheres is that the diameter difference is less than or equal to a preset value. q is the total number of pressure performance spheres in the diameter difference matrix, and b is the adjustment constant for equal influence.
[0013] In some embodiments disclosed in this invention, a tunnel construction safety intelligent management and control system based on dynamic sensing data includes: The first module is used to analyze historical tunnel data, including acquiring the surrounding rock pressure parameters, crown settlement parameters, and peripheral convergence parameters at different locations on the tunnel face at preset intervals as the tunnel advances. The crown settlement parameters include the unit settlement amount and cumulative settlement value per unit time, and the peripheral convergence parameters include the unit convergence amount and cumulative convergence amount per unit time. The second module is used to lock the tunnel face at preset intervals, determine the center point of each tunnel face, connect them to form a tunnel trend line, and construct a tunnel face representation component at the corresponding node of each tunnel face. The tunnel face representation component is used to represent the surrounding rock pressure parameters, arch settlement parameters and surrounding convergence parameters of the tunnel face. The combination of the tunnel trend line and several tunnel face representation components is identified as the tunnel perception representation model. The third module is used to adjust the tunnel perception performance model based on various parameters to obtain dynamic performance states, analyze several dynamic performance states, and identify the physical location segments and frame number segments with anomalies, and then identify the calibrated dynamic performance states as reference dynamic performance states. The fourth module is used to integrate several reference dynamic performance states to obtain a reference dynamic performance state analysis model, and to analyze real-time dynamic sensing data using the reference dynamic performance state analysis model.
[0014] This invention discloses an intelligent management and control method and system for tunnel construction safety based on dynamic sensing data, belonging to the field of tunnel management technology. The method includes analyzing historical tunnel data to acquire parameters such as face rock pressure, crown settlement, and surrounding convergence parameters at preset intervals; locking the face, determining the center point, connecting the points to form a tunnel trend line, and constructing a sensing model by combining performance components; obtaining a dynamic performance state based on parameter adjustments, analyzing and calibrating abnormal sections to obtain a reference dynamic performance state; and integrating the reference state into an analysis model for real-time data analysis. This invention employs vector lines, pressure spheres, surface fitting, finite element method, and Monte Carlo simulation to achieve anomaly risk quantification, feature retrieval analysis, and real-time comparison, thereby improving safety management efficiency.
[0015] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description
[0016] Figure 1 This is a flowchart illustrating the steps of the intelligent management and control method for tunnel construction safety based on dynamic sensing data disclosed in this embodiment of the invention. Detailed Implementation
[0017] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.
[0018] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings and specific embodiments. It should be understood that the preferred embodiments described herein are only for illustration and explanation of the present invention and should not be construed as limiting the scope of protection of the present invention. Those skilled in the art can make some non-essential improvements and adjustments based on the following content of the present invention. In the present invention, unless otherwise expressly specified and limited, the technical terms used in the present invention should have the ordinary meaning understood by those skilled in the art.
[0019] Example: This invention discloses an intelligent management and control method for tunnel construction safety based on dynamic sensing data. (See reference...) Figure 1 ,include: Step S100 involves analyzing historical tunnel data, including acquiring surrounding rock pressure parameters, crown settlement parameters, and peripheral convergence parameters at different locations on the tunnel face at preset intervals as the tunnel advances. The crown settlement parameters include the unit settlement amount and cumulative settlement value per unit time, while the peripheral convergence parameters include the unit convergence amount and cumulative convergence amount per unit time.
[0020] The principle of step S100 is to establish a basic database for tunnel construction safety assessment through systematic analysis of historical tunnel data. Specifically, as tunnel excavation progresses, key monitoring parameters are collected from different locations at the tunnel face (i.e., the tunnel excavation face) at preset time or distance intervals. These parameters include surrounding rock pressure (reflecting the compressive strength of the rock mass on the lining), crown settlement parameters (settlement per unit time is used to capture the instantaneous deformation rate, and cumulative settlement is used to assess the overall deformation trend), and perimeter convergence parameters (convergence per unit time reflects the speed of inward contraction of the surrounding rock, and cumulative convergence assesses the total deformation). The collection of these parameters is based on sensor networks (such as stress gauges and displacement gauges), aiming to capture the dynamic changes in the tunnel's geological environment and provide spatiotemporally continuous historical data support for subsequent model construction. This helps identify potential risk patterns and avoids analytical biases caused by relying on static data from a single point in time.
[0021] In step S200, the tunnel face is locked sequentially at preset intervals, and the center point of each tunnel face is determined. Lines are connected between them to form a tunnel trend line. At the corresponding position node of each tunnel face, a tunnel face representation component is constructed. The tunnel face representation component is used to represent the surrounding rock pressure parameters, arch settlement parameters, and surrounding convergence parameters of the tunnel face. The combination of the tunnel trend line and several tunnel face representation components is identified as the tunnel perception representation model.
[0022] The principle of step S200 is to transform the collected parameters into a visualized and structured representation model to achieve an intuitive representation of the tunnel's state. First, the current tunnel face is locked sequentially at preset intervals, its geometric center point is determined, and lines connecting these center points form a tunnel trend line. This simulates the actual path and tortuous shape of the tunnel, reflecting its three-dimensional spatial evolution. At each tunnel face node, a tunnel face representation component is constructed. This component uses abstract graphic elements (such as vector lines and spheres) to represent surrounding rock pressure, crown settlement, and surrounding convergence parameters. For example, vector length corresponds to deformation, and sphere size corresponds to pressure value, thus integrating multi-dimensional data into a unified visual component. Combining the tunnel trend line with these components constitutes the tunnel perception representation model. Its core lies in transforming scattered perception data into a coherent spatiotemporal framework, facilitating subsequent dynamic adjustments and anomaly detection, and improving data processing efficiency and interpretability.
[0023] Step S300: Based on various parameters, adjust the tunnel perception performance model to obtain dynamic performance state. Analyze several dynamic performance states and mark the physical location segments and frame number segments with anomalies. The marked dynamic performance state is then identified as the reference dynamic performance state.
[0024] The principle of step S300 lies in using historical parameters to dynamically adjust the perception model and perform anomaly analysis to extract reliable reference benchmarks. Specifically, based on parameters such as surrounding rock pressure, settlement, and convergence, the component states of the model are updated in real time (e.g., adjusting vector length and sphere diameter), generating a dynamic performance sequence. This sequence is similar to video frames, capturing the temporal evolution of the tunnel state. Then, multiple dynamic performances are analyzed, including marking vector groups with significant changes, fitting surfaces to identify anomaly patterns, and labeling risky physical location segments (specific tunnel sections) and frame segments (time windows). These labels are based on threshold judgment and geometric feature extraction. The labeled dynamic performances are recognized as reference dynamic performances. The purpose is to extract typical anomaly patterns from historical data as a knowledge base for subsequent real-time comparison and risk prediction, avoiding the blindness of purely data-driven approaches.
[0025] Step S400: Integrate several reference dynamic performance states to obtain a reference dynamic performance state analysis model, and use the reference dynamic performance state analysis model to analyze real-time dynamic sensing data.
[0026] The principle of step S400 lies in integrating multiple reference dynamic performance states into a comprehensive analysis model to achieve intelligent application of real-time data. First, through classification, clustering, or fusion algorithms (such as matching based on retrieval features), several reference dynamic performance states are integrated into a reference dynamic performance state analysis model. This model includes an anomaly pattern library, risk index mapping, and a comparison mechanism. Then, this model is used to analyze real-time dynamic sensing data (from sensors currently under construction), including constructing real-time performance states, extracting features (such as the length and average diameter of curved surface intersection edges), retrieving similar reference states, and calculating the degree of equivalence to output risk assessment and early warning signals. Its core lies in realizing the transfer of historical experience to real-time decision-making, improving the safety management level of tunnel construction through closed-loop feedback, reducing human intervention, and increasing response speed.
[0027] In some embodiments disclosed in this invention, the method for constructing a face representation model includes: Step S201: Determine the detection points set on the edge of the working face, and construct several representation vector lines based on the direction of each detection point relative to the center point. One end of the representation vector line coincides with the center point, and the other end points to the detection point. A pressure representation sphere is set at the end pointing to the detection point. The diameter of the pressure representation sphere is determined according to the surrounding rock pressure at the detection point, and the length of the representation vector line is determined according to the settlement or convergence.
[0028] The principle of step S201 lies in achieving an abstract and visual representation of the tunnel working face state through the geometric positioning and parameter mapping of preset detection points at the edge of the tunnel face. Specifically, multiple detection points on the edge of the tunnel face are first determined. These points are based on sensor placement (e.g., uniform distribution or key stress points). Then, with the center point of the tunnel face as the origin, a representation vector line is constructed according to the direction (e.g., radial angle) of each detection point relative to the center point. This vector line extends from the center point to the detection point, and its length ratio corresponds to the crown settlement or the surrounding convergence (unit settlement / convergence reflects immediate deformation, and cumulative value reflects the overall trend) to quantify the degree of deformation. A pressure representation sphere is set at the end of the vector line pointing to the detection point, with its diameter proportional to the surrounding rock pressure value (the greater the pressure, the larger the diameter), thus intuitively representing the local stress distribution. This vector-sphere combination construction method transforms multi-dimensional sensing data (such as pressure and deformation) into two-dimensional or three-dimensional graphic elements, facilitating subsequent model integration and dynamic analysis, improving the accuracy and visualization efficiency of anomaly detection, and avoiding the complexity of traditional table or curve representations.
[0029] In some embodiments disclosed in this invention, the method for analyzing dynamic behavior includes: Step S301 involves analyzing each dynamic performance frame in the dynamic performance state, including analyzing the length of the performance vector line in each dynamic performance frame and the diameter of each pressure performance sphere. The analysis method includes: Step S3011: Mark the vector lines whose length changes and whose change amplitude per unit time is greater than or equal to a preset value.
[0030] Step S3012: Combine several representation vector lines with similar positional relationships, including representation vector lines on the same face representation component and adjacent long face representation components, to form a marked representation vector line group.
[0031] Step S3013: Determine the ends of the representation vector lines in the marked representation vector line group, and perform surface fitting with the ends as boundaries to obtain the representation vector surface.
[0032] Step S3014: Save the performance vector surfaces corresponding to several consecutive frames of the dynamic performance state, as well as the corresponding pressure performance spheres, to obtain the performance state marked performance segments.
[0033] In some embodiments disclosed in this invention, the method for identifying abnormal physical segments and frame number segments includes: Step S302: The performance segment marked by the performance state is simulated and constructed in an equivalent manner to construct several extended performance segments. The combination of the performance segment marked by the performance state and the extended performance segments is denoted as the comprehensive performance segment set.
[0034] The principle of step S302 lies in achieving comprehensive coverage and enhanced robustness of anomalous regions through the equivalent simulation construction of the labeled performance segments. Specifically, the labeled performance segments (anomaly vector groups based on surface fitting) are first equivalently processed, that is, similar regions are simulated and constructed in terms of physical location and frame number dimensions. For example, neighboring virtual segments are generated through threshold similarity or clustering algorithms. If their parameter change trends are highly consistent with the labeled segments, several extended performance segments are constructed. This simulation construction avoids the limitations of isolated labeling and ensures that no anomalous patterns are missed. The original labeled segments and extended segments are combined into a comprehensive performance segment set. The purpose is to form a more complete anomalous dataset, which facilitates subsequent risk analysis and retrieval, improves the accuracy and continuity of identifying anomalous physical location segments (specific tunnel sections) and frame number segments (time series windows), and reduces misjudgments caused by data noise or local bias.
[0035] Step S303: Perform anomaly risk analysis on each performance segment in the comprehensive performance segment set, including simulating the surrounding rock pressure parameters, crown settlement parameters and surrounding convergence parameters corresponding to each performance segment, outputting the corresponding collapse probability, and determining the corresponding anomaly risk index based on the preset probability interval to which the collapse probability belongs. Step S304: Perform retrieval feature analysis on each performance segment in the performance comprehensive performance segment set, and classify the performance segments step by step based on the determined retrieval features, and determine the retrieval feature group corresponding to each category group to obtain the performance segment retrieval model.
[0036] In some embodiments disclosed in this invention, the method for simulating the surrounding rock pressure parameters, crown settlement parameters, and surrounding convergence parameters corresponding to each manifested segment includes: Step S3031: Construct a simulation environment, including extracting the corresponding surrounding rock pressure parameter sequence, crown settlement parameter sequence, and peripheral convergence parameter sequence based on the physical location and frame range of the performance segment. The surrounding rock pressure parameter sequence includes the time series of pressure values at each detection point, the crown settlement parameter sequence includes the settlement change sequence, and the peripheral convergence parameter sequence includes the convergence change sequence.
[0037] Step S3032: Preprocess the extracted parameter sequence, including filtering outliers in the sequence and smoothing it based on a preset time window to obtain a smoothed parameter sequence group.
[0038] Step S3033: Construct a tunnel collapse simulation model. The model includes a finite element analysis module and a probability statistics module. The finite element analysis module is used to simulate the mechanical effects of surrounding rock pressure, crown settlement and perimeter convergence on the tunnel structure. The probability statistics module is used to calculate conditional probabilities based on historical collapse data.
[0039] Step S3034: Input the smoothing parameter sequence group into the finite element analysis module to perform mechanical simulation, including calculating the stress field under the surrounding rock pressure distribution, the displacement field caused by the crown settlement and the deformation field caused by the surrounding convergence, and output the structural stability index under the simulation.
[0040] Step S3035: Input the structural stability index into the probability statistics module to calculate the collapse probability. This includes using the Monte Carlo simulation method to perform multiple iterative simulations of the random disturbance of the parameter sequence and to count the proportion of collapse events as the corresponding collapse probability.
[0041] Step S3036: Output the collapse probability and map it to the abnormal risk index according to the preset probability interval. The preset probability interval includes a low-risk interval, a medium-risk interval, and a high-risk interval, which correspond to different abnormal risk index values.
[0042] In some embodiments disclosed in this invention, a method for performing feature analysis on each performance segment in the performance comprehensive performance segment set includes: Step S3037: For the vertical direction of the tunnel, construct several levels of surface intercepting planes, analyze the surface intersection edges of each representation vector surface and surface intercepting plane, determine the length of the surface intersection edges and record it as the intersection edge length, record the intersection edge length corresponding to each level of surface intercepting plane, and obtain the intersection edge length sequence.
[0043] Step S3038: Take each cross edge length in the cross edge length sequence as a retrieval feature that is separated step by step, calculate the average diameter of the pressure representation sphere corresponding to the representation vector surface, and take the average diameter as the retrieval feature.
[0044] In some embodiments disclosed in this invention, the method for analyzing real-time dynamic sensing data using a reference dynamic performance analysis model includes: Step S401: Based on the real-time dynamic sensing data of the tunnel, construct the real-time performance segment, and perform feature analysis on the real-time performance segment to obtain the real-time cross-edge length sequence and the real-time average diameter of the pressure performance sphere.
[0045] Step S402: Based on the real-time cross-edge length sequence and the real-time average diameter, a search is performed to find several performance segments in the performance comprehensive performance segment set, which are recorded as the performance segments for comparison.
[0046] Step S403: Compare each comparison performance segment with the real-time performance segment to determine the degree of equivalence between the two, and based on the degree of equivalence, select a reference performance segment from the comparison performance segments.
[0047] In some embodiments disclosed in this invention, the method for calculating the degree of equivalence includes: Step S4031 involves performing several mutual attitude adjustments on the representation vector surfaces between the representation segments, including rotating and translating each pair of representation vector surfaces to minimize the geometric differences between the surfaces, calculating the proportion of overlapping areas of the adjusted surfaces as the attitude adjustment similarity, and selecting the optimal attitude adjustment method.
[0048] Step S4032: Extract feature points from the surface of the performance vector corresponding to the optimal posture adjustment method, including identifying key bending points, peak points and valley points on the surface, constructing a feature point matrix, matching and comparing the feature point matrix, calculating the distance between points, and obtaining the corresponding distance matrix between points.
[0049] Step S4033: Analyze the pressure representation spheres corresponding to the representation vector surface, determine the diameter difference between them, and obtain the diameter difference matrix.
[0050] Step S4034: Perform a comprehensive analysis of the point distance matrix and diameter difference matrix of each frame in the performance segment to determine the degree of equivalence.
[0051] In some embodiments disclosed in this invention, the expression for calculating the degree of equivalence is as follows: ; Where D represents the same degree. The distance matrix between points has an equal influence on the weights. The diameter difference matrix has equal influence weights. To preset the maximum distance between points as an equivalent parameter, To represent the number of matching points in the distance matrix corresponding to the i-th frame of the representation segment, the matching point is determined by the distance between points being less than or equal to a preset value, where h is the total number of points in the distance matrix. The number of matching spheres in the diameter difference matrix corresponding to the i-th frame of the performance segment is represented by q, and the method for judging matching spheres is that the diameter difference is less than or equal to a preset value. q is the total number of pressure performance spheres in the diameter difference matrix, and b is the adjustment constant for equal influence.
[0052] The distance matrix between points has an equal influence weight, a preset coefficient used to adjust the importance of distance matching between points in the overall equivalence calculation. The diameter difference matrix also has an equal influence weight, a preset coefficient used to adjust the importance of diameter difference matching in the overall equivalence calculation. The equivalence influence adjustment constant is a fixed bias value used to fine-tune the summation result, compensate for potential biases, or standardize the calculation.
[0053] In some embodiments disclosed in this invention, a tunnel construction safety intelligent management and control system based on dynamic sensing data includes: The first module is used to analyze historical tunnel data, including acquiring surrounding rock pressure parameters, crown settlement parameters, and peripheral convergence parameters at different locations on the tunnel face at preset intervals as the tunnel advances. The crown settlement parameters include the unit settlement amount and cumulative settlement value per unit time, and the peripheral convergence parameters include the unit convergence amount and cumulative convergence amount per unit time.
[0054] The second module is used to lock the tunnel face at preset intervals, determine the center point of each tunnel face, connect them to form a tunnel trend line, and construct a tunnel face representation component at the corresponding node of each tunnel face. The tunnel face representation component is used to represent the surrounding rock pressure parameters, arch settlement parameters, and surrounding convergence parameters of the tunnel face. The combination of the tunnel trend line and several tunnel face representation components is recognized as the tunnel perception representation model.
[0055] The third module is used to adjust the tunnel perception performance model based on various parameters to obtain dynamic performance states. Several dynamic performance states are analyzed, and the physical location segments and frame number segments with anomalies are identified. The identified dynamic performance states are then recognized as reference dynamic performance states.
[0056] The fourth module is used to integrate several reference dynamic performance states to obtain a reference dynamic performance state analysis model, and to analyze real-time dynamic sensing data using the reference dynamic performance state analysis model.
[0057] This invention discloses an intelligent management and control method and system for tunnel construction safety based on dynamic sensing data, belonging to the field of tunnel management technology. The method includes analyzing historical tunnel data to acquire parameters such as face rock pressure, crown settlement, and surrounding convergence parameters at preset intervals; locking the face, determining the center point, connecting the points to form a tunnel trend line, and constructing a sensing model by combining performance components; obtaining a dynamic performance state based on parameter adjustments, analyzing and calibrating abnormal sections to obtain a reference dynamic performance state; and integrating the reference state into an analysis model for real-time data analysis. This invention employs vector lines, pressure spheres, surface fitting, finite element method, and Monte Carlo simulation to achieve anomaly risk quantification, feature retrieval analysis, and real-time comparison, thereby improving safety management efficiency.
[0058] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented in hardware or by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, portable hard drive, etc.) and includes several instructions to cause a computer device (such as a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.
[0059] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.
Claims
1. A method for intelligent management and control of tunnel construction safety based on dynamic sensing data, characterized in that, include: Step S100: Analyze historical tunnel data, including obtaining surrounding rock pressure parameters, crown settlement parameters, and peripheral convergence parameters at different locations on the tunnel face at preset intervals as the tunnel advances. The crown settlement parameters include the unit settlement amount and cumulative settlement value per unit time, and the peripheral convergence parameters include the unit convergence amount and cumulative convergence amount per unit time. Step S200: Lock the tunnel face one by one according to the preset interval, determine the center point of each tunnel face, connect them with each other to form a tunnel trend line, and construct a tunnel face representation component at the corresponding position node of each tunnel face. The tunnel face representation component is used to represent the surrounding rock pressure parameters, arch settlement parameters and surrounding convergence parameters of the tunnel face. The combination of the tunnel trend line and several tunnel face representation components is identified as the tunnel perception representation model. Step S300: Based on various parameters, adjust the tunnel perception performance model to obtain dynamic performance state, analyze several dynamic performance states, and mark the physical location segments and frame number segments with anomalies, and identify the marked dynamic performance state as the reference dynamic performance state. Step S400: Integrate several reference dynamic performance states to obtain a reference dynamic performance state analysis model, and use the reference dynamic performance state analysis model to analyze real-time dynamic sensing data.
2. The intelligent management and control method for tunnel construction safety based on dynamic sensing data according to claim 1, characterized in that, Methods for constructing a face representation model include: Step S201: Determine the detection points set on the edge of the working face, and construct several representation vector lines based on the direction of each detection point relative to the center point. One end of the representation vector line coincides with the center point, and the other end points to the detection point. A pressure representation sphere is set at the end pointing to the detection point. The diameter of the pressure representation sphere is determined according to the surrounding rock pressure at the detection point, and the length of the representation vector line is determined according to the settlement or convergence.
3. The intelligent management and control method for tunnel construction safety based on dynamic sensing data according to claim 2, characterized in that, Methods for analyzing dynamic behavior include: Step S301 involves analyzing each dynamic performance frame in the dynamic performance state, including analyzing the length of the performance vector line in each dynamic performance frame and the diameter of each pressure performance sphere. The analysis method includes: Step S3011: Mark the vector lines whose length changes and whose change amplitude per unit time is greater than or equal to a preset value; Step S3012: Combine several representation vector lines with similar positional relationships, including representation vector lines on the same face representation component and adjacent long face representation components, to form a marked representation vector line group. Step S3013: Determine the ends of the representation vector lines in the marked representation vector line group, and perform surface fitting with the ends as boundaries to obtain the representation vector surface; Step S3014: Save the performance vector surfaces corresponding to several consecutive frames of the dynamic performance state, as well as the corresponding pressure performance spheres, to obtain the performance state marked performance segments.
4. The intelligent management and control method for tunnel construction safety based on dynamic sensing data according to claim 3, characterized in that, Methods for identifying abnormal physical and frame number segments include: Step S302: Equivalent expansion of the performance marker segment is performed to construct several performance extended segments. The combination of the performance marker segment and the performance extended segment is denoted as the performance comprehensive segment set. Step S303: Perform anomaly risk analysis on each performance segment in the comprehensive performance segment set, including simulating the surrounding rock pressure parameters, crown settlement parameters and surrounding convergence parameters corresponding to each performance segment, outputting the corresponding collapse probability, and determining the corresponding anomaly risk index based on the preset probability interval to which the collapse probability belongs. Step S304: Perform retrieval feature analysis on each performance segment in the performance comprehensive performance segment set, and classify the performance segments step by step based on the determined retrieval features, and determine the retrieval feature group corresponding to each category group to obtain the performance segment retrieval model.
5. The intelligent management and control method for tunnel construction safety based on dynamic sensing data according to claim 4, characterized in that, The methods for simulating the surrounding rock pressure parameters, crown settlement parameters, and surrounding convergence parameters corresponding to each manifested segment include: Step S3031: Construct a simulation environment, including extracting the corresponding surrounding rock pressure parameter sequence, crown settlement parameter sequence, and perimeter convergence parameter sequence based on the physical location and frame range of the performance segment. The surrounding rock pressure parameter sequence includes the time series of pressure values at each detection point, the crown settlement parameter sequence includes the settlement change sequence, and the perimeter convergence parameter sequence includes the convergence change sequence. Step S3032: Preprocess the extracted parameter sequence, including filtering outliers in the sequence and smoothing it based on a preset time window to obtain a smoothed parameter sequence group. Step S3033: Construct a tunnel collapse simulation model. The model includes a finite element analysis module and a probability statistics module. The finite element analysis module is used to simulate the mechanical effects of surrounding rock pressure, crown settlement and perimeter convergence on the tunnel structure. The probability statistics module is used to calculate conditional probabilities based on historical collapse data. Step S3034: Input the smoothing parameter sequence group into the finite element analysis module to perform mechanical simulation, including calculating the stress field under the surrounding rock pressure distribution, the displacement field caused by the crown settlement and the deformation field caused by the surrounding convergence, and output the structural stability index under the simulation. Step S3035: Input the structural stability index into the probability statistics module to calculate the collapse probability. This includes using the Monte Carlo simulation method to perform multiple iterative simulations of the random disturbance of the parameter sequence and to count the proportion of collapse events as the corresponding collapse probability. Step S3036: Output the collapse probability and map it to the abnormal risk index according to the preset probability interval. The preset probability interval includes a low-risk interval, a medium-risk interval, and a high-risk interval, which correspond to different abnormal risk index values.
6. The intelligent management and control method for tunnel construction safety based on dynamic sensing data according to claim 4, characterized in that, The methods for performing feature analysis on each performance segment in the comprehensive performance segment set include: Step S3037: For the vertical direction of the tunnel, construct several levels of surface intercepting planes, analyze the surface intersection edges of each representation vector surface and surface intercepting plane, determine the length of the surface intersection edges and record it as the intersection edge length, record the intersection edge length corresponding to each level of surface intercepting plane, and obtain the intersection edge length sequence. Step S3038: Take each cross edge length in the cross edge length sequence as a retrieval feature that is separated step by step, calculate the average diameter of the pressure representation sphere corresponding to the representation vector surface, and take the average diameter as the retrieval feature.
7. The intelligent management and control method for tunnel construction safety based on dynamic sensing data according to claim 4, characterized in that, Methods for analyzing real-time dynamic sensing data using a reference dynamic performance analysis model include: Step S401: Based on the real-time dynamic sensing data of the tunnel, construct the real-time performance segment, and perform feature analysis on the real-time performance segment to obtain the real-time cross-edge length sequence and the real-time average diameter of the pressure performance sphere. Step S402: Based on the real-time cross-edge length sequence and the real-time average diameter, a search is performed to find several performance segments in the performance comprehensive performance segment set, which are recorded as the performance segments for comparison. Step S403: Compare each comparison performance segment with the real-time performance segment to determine the degree of equivalence between the two, and based on the degree of equivalence, select a reference performance segment from the comparison performance segments.
8. The intelligent management and control method for tunnel construction safety based on dynamic sensing data according to claim 7, characterized in that, Methods for calculating equivalence include: Step S4031: Perform several mutual attitude adjustments on the representation vector surfaces between the representation segments, including rotating and translating each pair of representation vector surfaces to minimize the geometric differences between the surfaces, and calculate the ratio of the overlapping area of the adjusted surfaces as the attitude adjustment similarity, and select the optimal attitude adjustment method. Step S4032: Extract feature points from the surface of the performance vector corresponding to the optimal posture adjustment method, including identifying key bending points, peak points and valley points on the surface, constructing a feature point matrix, matching and comparing the feature point matrix, calculating the distance between points, and obtaining the corresponding distance matrix between points. Step S4033: Analyze the pressure-representing spheres corresponding to the surface of the representation vector and determine the diameter difference between them to obtain the diameter difference matrix; Step S4034: Perform a comprehensive analysis of the point distance matrix and diameter difference matrix of each frame in the performance segment to determine the degree of equivalence.
9. The intelligent management and control method for tunnel construction safety based on dynamic sensing data according to claim 8, characterized in that, The expression for calculating equivalence is: ； Where D represents the same degree. The distance matrix between points has an equal influence on the weights. The diameter difference matrix has equal influence weights. To preset the maximum distance between points as an equivalent parameter, To represent the number of matching points in the distance matrix corresponding to the i-th frame of the representation segment, the matching point is determined by the distance between points being less than or equal to a preset value, where h is the total number of points in the distance matrix. The number of matching spheres in the diameter difference matrix corresponding to the i-th frame of the performance segment is represented by q, and the method for judging matching spheres is that the diameter difference is less than or equal to a preset value. q is the total number of pressure performance spheres in the diameter difference matrix, and b is the adjustment constant for equal influence.
10. A tunnel construction safety intelligent management and control system based on dynamic sensing data, characterized in that: include: The first module is used to analyze historical tunnel data, including acquiring the surrounding rock pressure parameters, crown settlement parameters, and peripheral convergence parameters at different locations on the tunnel face at preset intervals as the tunnel advances. The crown settlement parameters include the unit settlement amount and cumulative settlement value per unit time, and the peripheral convergence parameters include the unit convergence amount and cumulative convergence amount per unit time. The second module is used to lock the tunnel face at preset intervals, determine the center point of each tunnel face, connect them to form a tunnel trend line, and construct a tunnel face representation component at the corresponding node of each tunnel face. The tunnel face representation component is used to represent the surrounding rock pressure parameters, arch settlement parameters and surrounding convergence parameters of the tunnel face. The combination of the tunnel trend line and several tunnel face representation components is identified as the tunnel perception representation model. The third module is used to adjust the tunnel perception performance model based on various parameters to obtain dynamic performance states, analyze several dynamic performance states, and identify the physical location segments and frame number segments with anomalies, and then identify the calibrated dynamic performance states as reference dynamic performance states. The fourth module is used to integrate several reference dynamic performance states to obtain a reference dynamic performance state analysis model, and to analyze real-time dynamic sensing data using the reference dynamic performance state analysis model.