Dynamic risk assessment method for large-diameter shield tunnel karst cave

Abstract

The invention belongs to the technical field of tunnel engineering, and discloses a large-diameter shield tunnel karst cave dynamic risk assessment method, which comprises the following specific steps: step 1, constructing a basic data acquisition and assessment model before construction according to tunnel construction design specifications, and acquiring karst geological data within a range of 50 meters from the axis of a tunnel and the periphery of the tunnel; an evaluation foundation is laid by building an initial risk evaluation model before construction, multi-dimensional data are collected and evaluation indexes and weights are dynamically updated during construction, real-time calculation and grade judgment of risks are achieved, meanwhile, a risk evaluation result is closely linked with countermeasures, and adaptive management and control means are rapidly matched according to different risk grades. And a closed loop is formed by combining data recording and model optimization after construction, so that the capture capability and evaluation accuracy of karst cave risk dynamic evolution are effectively improved, the collaboration of risk evaluation and response measures is enhanced, and a guarantee is provided for construction safety of a large-diameter shield tunnel under karst geological conditions.

Description

Technical Field

[0001] This invention belongs to the field of tunnel engineering technology, specifically a dynamic risk assessment method for karst caves in large-diameter shield tunnels. Background Technology

[0002] In the construction of major projects such as urban rail transit and cross-river and cross-sea passages, large-diameter shield tunnels are being used more and more widely due to their high construction efficiency and minimal disturbance to the surrounding environment.

[0003] However, many tunnel projects need to traverse karst-developed areas. Karst caves are characterized by their strong concealment and complex development patterns. If the risks of karst caves are not effectively managed during construction, they can easily lead to safety accidents such as tunnel collapses, water inrushes, and sand inrushes. Therefore, risk assessment of karst caves is a critical link in the construction safety of large-diameter shield tunnels. Currently, there are risk assessment technologies for karst caves in shield tunnels in the industry, but these technologies still have significant shortcomings in practical applications. Existing assessments are mostly based on pre-construction survey data to conduct static assessments. After determining the risk level, they rarely incorporate actual geological changes and construction parameters during construction for dynamic adjustments, making it difficult for the assessment results to adapt to real-time risk conditions. At the same time, the integration and utilization of real-time karst cave detection information, shield machine operation data, and environmental monitoring data generated during construction is low, making it impossible to capture the dynamic evolution of karst cave risks in a timely manner. Furthermore, there is a lack of close linkage between risk assessment and response measures, making it difficult to quickly match effective control measures according to real-time changes in risks. The overall process fails to form a complete closed loop of dynamic risk perception, assessment, and response, which is insufficient to meet the actual needs of safety management in the construction of large-diameter shield tunnels. Therefore, improvements are needed. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a dynamic risk assessment method for karst caves in large-diameter shield tunnels, which has the advantage of dynamic assessment.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for dynamic risk assessment of karst caves in large-diameter shield tunnels, the specific steps of which are as follows:

[0006] Step 1: Pre-construction basic data collection and assessment model building: Collect karst geology, tunnel design and construction parameters, and build an initial risk assessment model containing core indicators and initial weights;

[0007] In accordance with tunnel construction design specifications, karst geological data were collected within a 50-meter radius of the tunnel axis, including the distribution range of soluble rocks, rock layer thickness, rock layer dip angle, groundwater level, groundwater flow velocity, and groundwater flow direction. Drilling exploration was conducted using a 150 mm borehole, with one borehole every 20 meters to obtain data on karst cave development within the boreholes, including cave depth, transverse diameter, longitudinal length, type of infill material, and density of the infill material, calculated by weight, with values ​​ranging from 0-100%. Design and construction parameters for large-diameter shield tunnels were collected, including tunnel inner diameter, tunnel outer diameter, tunnel depth, shield machine model, shield machine design parameters, and segment design parameters. An initial risk assessment model was established, determining core assessment indicators and initial weights: core indicators included the minimum distance between the karst cave and the tunnel structure, and the volume of the karst cave. Stability of karst cave filling: For mud-filled karst caves, the water content of the silt is classified as follows: ≤30% is stable, 30%-60% is relatively stable, and >60% is unstable. For water-filled karst caves, the water level difference between the cave and the tunnel is classified as follows: ≤5 meters is stable, 5-10 meters is relatively stable, and >10 meters is unstable. Empty karst caves are directly classified as unstable. Ground subsidence warning values: ≤3 mm for residential areas and ≤5 mm for municipal road areas. Tunnel boring machine (TBM) propulsion parameter deviation values: Based on accident case data and normal construction data from over 100 sets of large-diameter TBM tunnel projects in karst areas, the initial weights of each indicator were calculated using a data fitting method. The weights for the minimum distance between the karst cave and the tunnel are 0.25, the karst cave volume is 0.2, the filling stability is 0.2, the ground subsidence warning value is 0.15, and the TBM propulsion parameter deviation value is 0.2.

[0008] Step 2: Initial risk level determination before construction: Substitute the data and calculate the initial risk value using the weighted summation method, classify it into low / medium / high risk levels, and formulate corresponding initial response preparation measures;

[0009] Substitute the basic geological data and tunnel parameters collected in Step 1 into the initial risk assessment model, and calculate the initial risk value using a weighted summation method: First, convert each indicator into a quantitative score according to the "risk level - numerical value" correspondence: minimum distance between the karst cave and the tunnel <2 meters is 100 points, 2-5 meters is 70 points, 5-10 meters is 40 points, and >10 meters is 10 points; karst cave volume >100 cubic meters is 100 points, 50-100 cubic meters is 70 points, 10- 50 cubic meters = 40 points, <10 cubic meters = 10 points; Unstable infill material = 100 points, relatively stable = 50 points, stable = 10 points; Ground settlement warning value: Actual monitored value ≥ warning value = 100 points, 0.5 × warning value - warning value = 50 points, <0.5 × warning value = 10 points; Tunnel boring machine propulsion parameter deviation >20% = 100 points, 10%-20% = 50 points, <10% = 10 points.

[0010] The initial risk value is obtained by multiplying the quantitative scores of each indicator by their corresponding initial weights and summing the results. The score range is 0-100 points, and the initial risk levels are divided as follows: 0-30 points is low risk level, 31-60 points is medium risk level, and 61-100 points is high risk level. Initial response and preparation measures are formulated for different levels. For low risk level, settlement monitoring instruments and water level monitoring instruments are prepared. For medium risk level, ground-penetrating radar is added to the low risk equipment. For high risk level, grouting pumps, grouting materials, and quick-setting concrete are prepared in addition to the medium risk equipment.

[0011] Step 3: Real-time data acquisition during construction: Collect real-time data on shield machine operating parameters, karst cave detection data, water level, ground settlement, and segment stress at the prescribed frequency;

[0012] The tunnel boring machine (TBM) utilizes its built-in sensor system to collect operating parameters every minute, including actual thrust, torque, propulsion speed, excavated soil volume, cutterhead rotation speed, and TBM attitude parameters. This data is transmitted in real-time to the tunnel monitoring center database. A 250 MHz ground-penetrating radar is used to conduct supplementary karst cave detection in front of the TBM cutterhead and on both sides of the tunnel. Detection is performed every 5 meters of tunnel construction length, acquiring the real-time location, size, and filling status of the caves. The detection data is then processed by the radar. After being processed by the software, the data is transmitted to the monitoring center database every 5 minutes. A water level monitoring point is set up every 10 meters along the tunnel axis, and a water level sensor is used to collect the water level height in the tunnel every 30 minutes. A settlement monitoring point is set up every 20 meters along the ground axis above the tunnel, and a high-precision settlement monitoring instrument is used to collect the ground settlement every hour. A stress monitoring point is set up every 5 rings at the tunnel segment circumferential joints, and a stress sensor is used to collect the stress value of the segment every 2 hours. All monitoring data is uploaded to the monitoring center database in real time.

[0013] Step 4: Real-time data processing and evaluation indicator update: Preprocess real-time data, update core evaluation indicators, and dynamically adjust indicator weights;

[0014] The real-time data collected in step three is preprocessed. Values ​​exceeding the mean ± 3 standard deviations are identified as outliers. Outliers caused by equipment malfunctions are removed. A moving average method with a window size of 5 collection periods is then used to smooth the outlier-removed data, resulting in a stable real-time data sequence. Based on the processed real-time data, the core evaluation indicators from step one are updated: the minimum distance between the karst cave and the tunnel structure is recalculated based on the real-time location detected by ground-penetrating radar; the karst cave volume is recalculated based on the real-time dimensions; the stability of the karst cave filling is reassessed based on the real-time condition; and the ground subsidence warning value is adjusted based on the real-time ground subsidence and the warning value. Ratio updates: The deviation values ​​of the tunnel boring machine's propulsion parameters are updated based on the real-time difference between the actual parameters and the rated parameters. The weights of the indicators are adjusted according to the risk correlation reflected by the real-time data: when the ground settlement reaches 80% or more of the warning value, the weight of the ground settlement warning value is adjusted from 0.15 to 0.3; when the deviation value of the tunnel boring machine's propulsion parameters reaches 15% or more of the rated value, the weight of the propulsion parameter deviation value is adjusted from 0.2 to 0.3; when the minimum distance between the karst cave and the tunnel is <3 meters, the weight of this indicator is adjusted from 0.25 to 0.35; in other cases, the initial weights are maintained, and the weight adjustments ensure that the sum of the weights of all indicators is 1.

[0015] Step 5: Dynamic Risk Value Calculation and Level Determination: Dynamic Risk Value Calculation and Level Determination: Substitute the updated indicators and weights, calculate the real-time risk value every 1 minute, and lock the level and provide an audio-visual prompt after 5 consecutive minutes of stability;

[0016] The updated evaluation indicators from step four are quantified and scored using the same quantification standards as in step two. These, along with the dynamically adjusted weights, are then substituted into the evaluation model. A weighted summation method is used to calculate the real-time risk value, with the calculation frequency matching the tunnel boring machine parameter acquisition frequency, maintained at once every minute. The real-time risk value is simultaneously displayed on the monitoring center console. The current risk level is determined according to the level classification standards of step two. When the real-time risk value remains within a certain level range for 5 consecutive minutes, the system automatically locks that level and issues an audible and visual alert. If the real-time risk value fluctuates across levels within 5 minutes, the system does not lock the level temporarily, but continues to calculate and display the fluctuation trend until it stabilizes at a certain level for 5 consecutive minutes.

[0017] Step Six: Dynamic Adjustment and Implementation of Risk Response Measures: Implement corresponding response measures such as shield machine parameter control, monitoring frequency adjustment, and grouting filling according to low / medium / high risk levels;

[0018] When the system locks in a low-risk level, maintain the tunnel boring machine's normal advance parameters: advance speed 20-30 mm / min, cutterhead speed 2-3 rpm, and excavation volume controlled within ±5% of the design value. Continue collecting various data at the original frequency, summarizing and analyzing the monitoring data daily. After confirming no abnormal changes in risk indicators, proceed with construction as planned the following day. When the system locks in a medium-risk level, immediately reduce the tunnel boring machine's advance speed to 10-15 mm / min, cutterhead speed to 1-2 rpm, and control the excavation volume to 80%-90% of the design value. Increase the frequency of ground-penetrating radar detection to once every 2 meters and the frequency of ground settlement monitoring to once every 30 minutes. Assign dedicated personnel to manually check the tunnel boring machine's attitude and segment stress data every hour. If the segment stress value is found to exceed 90% of the design value, immediately suspend advance and investigate the cause. Only after the data returns to normal can the tunnel boring machine (TBM) continue its advance. When the system locks in a high-risk level, immediately stop the TBM's advance, shut down the cutterhead drive system, and activate the emergency drainage equipment inside the tunnel with a drainage capacity of ≥50 cubic meters per hour. Start the grouting pump, calculate the grouting volume based on 1.2 times the volume of the karst cave, and inject cement grout into the karst cave through the grouting holes on the TBM cutterhead. During the grouting process, collect the pressure data inside the karst cave every 10 minutes. Stop grouting when the pressure stabilizes at 0.3-0.5 MPa. After grouting, let it stand for 24 hours, then use ground-penetrating radar to detect the filling condition of the karst cave. After confirming that the filling density is ≥90%, recalculate the real-time risk value. If the risk value drops below 30 points, resume the advance according to the low-risk level parameters, with an initial advance speed of 5 mm / min, gradually increasing to the normal speed. If the risk value is still higher than 30 points, repeat the grouting process until the risk level drops to low risk.

[0019] Step 7: Data Recording and Evaluation Model Optimization: Record data daily to form a daily report archive. After each 100-meter tunnel construction is completed, summarize the data, calibrate the weights, supplement the quantitative standards, and optimize the evaluation model.

[0020] After each day's construction is completed, detailed records are kept of the day's basic data, real-time collected data, data processing results, risk value calculation process, risk level determination results, and the implementation and effects of countermeasures, i.e., changes in risk values ​​and settlement before and after grouting. This forms a daily dynamic risk assessment report. The report includes data tables, risk level change curves, photos of countermeasure implementation, and text descriptions. All daily reports are archived and stored according to the construction date. After every 100 meters of tunnel construction is completed, all daily report data for that section are summarized, and regression analysis is used to recalibrate the weights of the assessment indicators. Indicators with weight deviations exceeding 5% are adjusted. Based on the type of karst cave encountered in that section, specific quantitative standards for that type of karst cave are supplemented. The optimized model is used for dynamic risk assessment of subsequent sections to ensure that the model is adapted to the karst geological characteristics of different construction sections.

[0021] Preferably, the drilling equipment in step one should be a hydraulic core drilling rig, the verticality deviation of the drilling rig should be controlled within ≤0.5%, a core sample should be taken every 5 meters during the drilling process, and the core recovery rate should be ≥90%.

[0022] Preferably, the accident case data and normal construction data of the large-diameter shield tunnel project in the karst area mentioned in step one should be selected from large-diameter shield tunnel projects in the karst area completed within the past 10 years with a geological condition similarity of ≥80%, to ensure the reference value of the data.

[0023] Preferably, the shield machine sensors mentioned in step three need to be calibrated once a month, with the calibration error controlled within ±2%; before the ground-penetrating radar detection, it needs to be debugged in a simulated karst cave environment of known size to ensure that the detection error is ≤5%, and the distance between the radar antenna and the tunnel wall remains stable during detection, with a deviation of ≤2 cm.

[0024] Preferably, when applying the processed real-time data described in step four, the collected data must first undergo a normality test, and the criterion for removing outliers can only be used if the pass rate of the test is ≥95%.

[0025] Preferably, after the weight adjustment in step four, it needs to be verified by more than 5 sets of historical risk case data. If the verification accuracy rate is ≥90%, the adjustment is confirmed to be effective; otherwise, the weight calculation method should be re-optimized.

[0026] Preferably, the cement used for the grouting material in step six should be 42.5 grade ordinary Portland cement, and the admixture should be a high-efficiency water-reducing agent; the geological radar detection after grouting should cover the entire area of ​​the karst cave, with the detection point spacing ≤ 1 meter, to ensure comprehensive verification of the filling density.

[0027] Preferably, the audio-visual prompts in step five are: a green light for low risk and a prompt tone at a frequency of 1 time per minute; a yellow light for medium risk and a prompt tone at a frequency of 2 times per minute; and a red light for high risk and a prompt tone at a frequency of 3 times per minute.

[0028] Preferably, the water level sensor and stress sensor mentioned in step three need to be waterproofed before installation, with a waterproof rating of IP68. The sensor connection cable should be armored cable to avoid data transmission interruption due to damage during construction. When laying the cable, it should be fixed inside the tunnel segment with a fixing spacing of ≤1 meter.

[0029] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0030] By establishing an initial risk assessment model before construction to lay the foundation for assessment, and collecting multi-dimensional data in real time during construction and dynamically updating assessment indicators and weights, the risk can be calculated and its level determined in real time. At the same time, the risk assessment results are closely linked with the countermeasures, and appropriate control measures are quickly matched according to different risk levels. Combined with data recording and model optimization after construction, a complete closed loop is formed, which effectively improves the ability to capture the dynamic evolution of karst cave risks and the accuracy of assessment, enhances the synergy between risk assessment and countermeasures, provides more reliable protection for the construction safety of large-diameter shield tunnels under karst geological conditions, and meets the actual safety management needs of the project. Attached Figure Description

[0031] Figure 1 This is a flowchart of the dynamic risk assessment method for karst caves in large-diameter shield tunnels according to the present invention. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] like Figure 1 As shown in the figure, this invention provides a method for dynamic risk assessment of karst caves in large-diameter shield tunnels, with the following specific steps:

[0034] Step 1: Pre-construction basic data collection and assessment model building: Collect karst geology, tunnel design and construction parameters, and build an initial risk assessment model containing core indicators and initial weights;

[0035] In accordance with tunnel construction design specifications, karst geological data were collected within a 50-meter radius of the tunnel axis, including the distribution range of soluble rocks, rock layer thickness, rock layer dip angle, groundwater level, groundwater flow velocity, and groundwater flow direction. Drilling exploration was conducted using a 150 mm borehole, with one borehole every 20 meters to obtain data on karst cave development within the boreholes, including cave depth, transverse diameter, longitudinal length, type of infill material, and density of the infill material, calculated by weight, with values ​​ranging from 0-100%. Design and construction parameters for large-diameter shield tunnels were collected, including tunnel inner diameter, tunnel outer diameter, tunnel depth, shield machine model, shield machine design parameters, and segment design parameters. An initial risk assessment model was established, determining core assessment indicators and initial weights: core indicators included the minimum distance between the karst cave and the tunnel structure, and the volume of the karst cave. Stability of karst cave filling: For mud-filled karst caves, the water content of the silt is classified as follows: ≤30% is stable, 30%-60% is relatively stable, and >60% is unstable. For water-filled karst caves, the water level difference between the cave and the tunnel is classified as follows: ≤5 meters is stable, 5-10 meters is relatively stable, and >10 meters is unstable. Empty karst caves are directly classified as unstable. Ground subsidence warning values: ≤3 mm for residential areas and ≤5 mm for municipal road areas. Tunnel boring machine (TBM) propulsion parameter deviation values: Based on accident case data and normal construction data from over 100 sets of large-diameter TBM tunnel projects in karst areas, the initial weights of each indicator were calculated using a data fitting method. The weights for the minimum distance between the karst cave and the tunnel are 0.25, the karst cave volume is 0.2, the filling stability is 0.2, the ground subsidence warning value is 0.15, and the TBM propulsion parameter deviation value is 0.2.

[0036] Step 2: Initial risk level determination before construction: Substitute the data and calculate the initial risk value using the weighted summation method, classify it into low / medium / high risk levels, and formulate corresponding initial response preparation measures;

[0037] Substitute the basic geological data and tunnel parameters collected in Step 1 into the initial risk assessment model, and calculate the initial risk value using a weighted summation method: First, convert each indicator into a quantitative score according to the "risk level - numerical value" correspondence: minimum distance between the karst cave and the tunnel <2 meters is 100 points, 2-5 meters is 70 points, 5-10 meters is 40 points, and >10 meters is 10 points; karst cave volume >100 cubic meters is 100 points, 50-100 cubic meters is 70 points, 10- 50 cubic meters = 40 points, <10 cubic meters = 10 points; Unstable infill material = 100 points, relatively stable = 50 points, stable = 10 points; Ground settlement warning value: Actual monitored value ≥ warning value = 100 points, 0.5 × warning value - warning value = 50 points, <0.5 × warning value = 10 points; Tunnel boring machine propulsion parameter deviation >20% = 100 points, 10%-20% = 50 points, <10% = 10 points.

[0038] The initial risk value is obtained by multiplying the quantitative scores of each indicator by their corresponding initial weights and summing the results. The score range is 0-100 points, and the initial risk levels are divided as follows: 0-30 points is low risk level, 31-60 points is medium risk level, and 61-100 points is high risk level. Initial response and preparation measures are formulated for different levels. For low risk level, settlement monitoring instruments and water level monitoring instruments are prepared. For medium risk level, ground-penetrating radar is added to the low risk equipment. For high risk level, grouting pumps, grouting materials, and quick-setting concrete are prepared in addition to the medium risk equipment.

[0039] Step 3: Real-time data acquisition during construction: Collect real-time data on shield machine operating parameters, karst cave detection data, water level, ground settlement, and segment stress at the prescribed frequency;

[0040] The tunnel boring machine (TBM) utilizes its built-in sensor system to collect operating parameters every minute, including actual thrust, torque, propulsion speed, excavated soil volume, cutterhead rotation speed, and TBM attitude parameters. This data is transmitted in real-time to the tunnel monitoring center database. A 250 MHz ground-penetrating radar is used to conduct supplementary karst cave detection in front of the TBM cutterhead and on both sides of the tunnel. Detection is performed every 5 meters of tunnel construction length, acquiring the real-time location, size, and filling status of the caves. The detection data is then processed by the radar. After being processed by the software, the data is transmitted to the monitoring center database every 5 minutes. A water level monitoring point is set up every 10 meters along the tunnel axis, and a water level sensor is used to collect the water level height in the tunnel every 30 minutes. A settlement monitoring point is set up every 20 meters along the ground axis above the tunnel, and a high-precision settlement monitoring instrument is used to collect the ground settlement every hour. A stress monitoring point is set up every 5 rings at the tunnel segment circumferential joints, and a stress sensor is used to collect the stress value of the segment every 2 hours. All monitoring data is uploaded to the monitoring center database in real time.

[0041] Step 4: Real-time data processing and evaluation indicator update: Preprocess real-time data, update core evaluation indicators, and dynamically adjust indicator weights;

[0042] The real-time data collected in step three is preprocessed. Values ​​exceeding the mean ± 3 standard deviations are identified as outliers. Outliers caused by equipment malfunctions are removed. A moving average method with a window size of 5 collection periods is then used to smooth the outlier-removed data, resulting in a stable real-time data sequence. Based on the processed real-time data, the core evaluation indicators from step one are updated: the minimum distance between the karst cave and the tunnel structure is recalculated based on the real-time location detected by ground-penetrating radar; the karst cave volume is recalculated based on the real-time dimensions; the stability of the karst cave filling is reassessed based on the real-time condition; and the ground subsidence warning value is adjusted based on the real-time ground subsidence and the warning value. Ratio updates: The deviation values ​​of the tunnel boring machine's propulsion parameters are updated based on the real-time difference between the actual parameters and the rated parameters. The weights of the indicators are adjusted according to the risk correlation reflected by the real-time data: when the ground settlement reaches 80% or more of the warning value, the weight of the ground settlement warning value is adjusted from 0.15 to 0.3; when the deviation value of the tunnel boring machine's propulsion parameters reaches 15% or more of the rated value, the weight of the propulsion parameter deviation value is adjusted from 0.2 to 0.3; when the minimum distance between the karst cave and the tunnel is <3 meters, the weight of this indicator is adjusted from 0.25 to 0.35; in other cases, the initial weights are maintained, and the weight adjustments ensure that the sum of the weights of all indicators is 1.

[0043] Step 5: Dynamic Risk Value Calculation and Level Determination: Substitute the updated indicators and weights, calculate the real-time risk value every minute, and lock the level and provide an audio-visual alert after 5 consecutive minutes of stability;

[0044] The updated evaluation indicators from step four are quantified and scored using the same quantification standards as in step two. These, along with the dynamically adjusted weights, are then substituted into the evaluation model. A weighted summation method is used to calculate the real-time risk value, with the calculation frequency matching the tunnel boring machine parameter acquisition frequency, maintained at once every minute. The real-time risk value is simultaneously displayed on the monitoring center console. The current risk level is determined according to the level classification standards of step two. When the real-time risk value remains within a certain level range for 5 consecutive minutes, the system automatically locks that level and issues an audible and visual alert. If the real-time risk value fluctuates across levels within 5 minutes, the system does not lock the level temporarily, but continues to calculate and display the fluctuation trend until it stabilizes at a certain level for 5 consecutive minutes.

[0045] Step Six: Dynamic Adjustment and Implementation of Risk Response Measures: Implement corresponding response measures such as shield machine parameter control, monitoring frequency adjustment, and grouting filling according to low / medium / high risk levels;

[0046] When the system locks in a low-risk level, maintain the tunnel boring machine's normal advance parameters: advance speed 20-30 mm / min, cutterhead speed 2-3 rpm, and excavation volume controlled within ±5% of the design value. Continue collecting various data at the original frequency, summarizing and analyzing the monitoring data daily. After confirming no abnormal changes in risk indicators, proceed with construction as planned the following day. When the system locks in a medium-risk level, immediately reduce the tunnel boring machine's advance speed to 10-15 mm / min, cutterhead speed to 1-2 rpm, and control the excavation volume to 80%-90% of the design value. Increase the frequency of ground-penetrating radar detection to once every 2 meters and the frequency of ground settlement monitoring to once every 30 minutes. Assign dedicated personnel to manually check the tunnel boring machine's attitude and segment stress data every hour. If the segment stress value is found to exceed 90% of the design value, immediately suspend advance and investigate the cause. Only after the data returns to normal can the tunnel boring machine (TBM) continue its advance. When the system locks in a high-risk level, immediately stop the TBM's advance, shut down the cutterhead drive system, and activate the emergency drainage equipment inside the tunnel with a drainage capacity of ≥50 cubic meters per hour. Start the grouting pump, calculate the grouting volume based on 1.2 times the volume of the karst cave, and inject cement grout into the karst cave through the grouting holes on the TBM cutterhead. During the grouting process, collect the pressure data inside the karst cave every 10 minutes. Stop grouting when the pressure stabilizes at 0.3-0.5 MPa. After grouting, let it stand for 24 hours, then use ground-penetrating radar to detect the filling condition of the karst cave. After confirming that the filling density is ≥90%, recalculate the real-time risk value. If the risk value drops below 30 points, resume the advance according to the low-risk level parameters, with an initial advance speed of 5 mm / min, gradually increasing to the normal speed. If the risk value is still higher than 30 points, repeat the grouting process until the risk level drops to low risk.

[0047] Step 7: Data Recording and Evaluation Model Optimization: Record data daily to form a daily report archive. After each 100-meter tunnel construction is completed, summarize the data, calibrate the weights, supplement the quantitative standards, and optimize the evaluation model.

[0048] After each day's construction is completed, detailed records are kept of the day's basic data, real-time collected data, data processing results, risk value calculation process, risk level determination results, and the implementation and effects of countermeasures, i.e., changes in risk values ​​and settlement before and after grouting. This forms a daily dynamic risk assessment report. The report includes data tables, risk level change curves, photos of countermeasure implementation, and text descriptions. All daily reports are archived and stored according to the construction date. After every 100 meters of tunnel construction is completed, all daily report data for that section are summarized, and regression analysis is used to recalibrate the weights of the assessment indicators. Indicators with weight deviations exceeding 5% are adjusted. Based on the type of karst cave encountered in that section, specific quantitative standards for that type of karst cave are supplemented. The optimized model is used for dynamic risk assessment of subsequent sections to ensure that the model is adapted to the karst geological characteristics of different construction sections.

[0049] Before construction, karst geology and tunnel design and construction parameters are collected to build an initial risk assessment model containing core evaluation indicators and initial weights. Initial risk values ​​are calculated and classified, and corresponding initial response measures are formulated. During construction, multiple types of data are collected and preprocessed in real time. Evaluation indicators are updated and weights are dynamically adjusted. Real-time risk values ​​are calculated periodically to determine the risk level and trigger alerts. Different response measures are implemented according to the risk level. After construction, daily data is recorded and archived. Data is summarized and the model is optimized after each certain tunnel construction length is completed to adapt to subsequent segment assessments, achieving dynamic and effective control of karst cave risks in this type of tunnel. The minimum distance between the cave and the tunnel is quantified based on the "proximity construction risk theory." The closer the cave is to the tunnel structure, the higher the probability of cave collapse causing a sudden change in tunnel load, and the greater the additional impact on the tunnel structure. The greater the stress, the more it is considered in conjunction with the "risk classification of proximity distance" logic in the "Technical Specification for Close-proximity Construction of Shield Tunnels"; the quantification of karst cave volume is based on the engineering practice principle that "the larger the karst cave volume, the wider its potential impact on tunnel construction and the weaker its risk bearing capacity"; the stability of karst cave filling: the quantification of mud-filled karst caves is based on the principle of "negative correlation between water content and strength of cohesive soil" in geotechnical mechanics; the quantification of water-filled karst caves is based on the "Design Code for Permeability Grade of Concrete Structures"; empty karst caves are prone to collapse directly during construction due to the lack of filling material to provide support; the quantification of ground settlement warning value is based on the "Technical Specification for Monitoring of Urban Rail Transit Engineering"; the quantification of shield machine propulsion parameter deviation value is based on the shield machine design manual, engineering construction experience, and the "Code for Construction and Acceptance of Shield Tunnels" (GB50446-2017).

[0050] In step one, the drilling equipment must be a hydraulic core drilling rig, the verticality deviation of the drilling rig must be controlled within ≤0.5%, and a core sample must be taken every 5 meters during the drilling process, with a core recovery rate of ≥90%.

[0051] By limiting the drilling equipment to hydraulic core drilling rigs, the borehole verticality deviation to ≤0.5%, and the core sampling rate to ≥90% every 5 meters, the accuracy of borehole exploration and the integrity of core samples can be ensured. This avoids data distortion caused by improper drilling equipment or non-standard operation, provides reliable data support for basic geological data collection, and ensures the accuracy of subsequent initial risk assessment model construction.

[0052] In step one, the accident case data and normal construction data for large-diameter shield tunnel projects in karst areas should be selected from large-diameter shield tunnel projects in karst areas completed within the past 10 years with a geological condition similarity of ≥80%, to ensure the reference value of the data.

[0053] By selecting projects completed within the last 10 years with a geological similarity of ≥80%, the geological correlation between historical data and current large-diameter shield tunnel projects can be ensured, reducing the impact of data deviation on the initial weight calculation, improving the rationality of the initial risk assessment model weight setting, and making the initial risk assessment more in line with actual engineering scenarios.

[0054] In step three, the shield machine sensors need to be calibrated monthly, with the calibration error controlled within ±2%. Before the ground-penetrating radar is used for detection, it needs to be debugged in a simulated karst cave environment of known size to ensure that the detection error is ≤5%. During detection, the distance between the radar antenna and the tunnel wall should remain stable with a deviation of ≤2 cm.

[0055] Calibrating the sensors ensures the accuracy of real-time data acquisition equipment and prevents distortion of tunnel boring machine operating parameters and karst cave detection data due to equipment errors.

[0056] When applying the processed real-time data described in step four, the collected data must first undergo a normality test. Only when the pass rate of the test is ≥95% can the criterion be used to remove outliers.

[0057] By verifying the data, we can ensure its stability and accuracy.

[0058] In step four, the weight adjustment needs to be verified using more than five sets of historical risk case data. If the verification accuracy is ≥90%, the adjustment is confirmed to be effective; otherwise, the weight calculation method needs to be re-optimized.

[0059] The weights are validated after adjustment to ensure the effectiveness of the adjustment, avoid mismatch between the adjusted weights and the actual risk correlation, ensure that the updated assessment indicator weights can truly reflect the degree of risk impact, and improve the credibility of the dynamic risk assessment model.

[0060] In step six, the cement used for grouting materials must be 42.5 grade ordinary Portland cement, and the admixture must be a high-efficiency water-reducing agent. After grouting is completed, the geological radar detection must cover the entire area of ​​the karst cave, with the detection point spacing ≤ 1 meter, to ensure comprehensive verification of the filling density.

[0061] By conducting detection after grouting is completed, with the detection point spacing ≤1 meter, the density of the karst cave filling can be accurately confirmed, avoiding the risk of residual hazards caused by substandard grouting.

[0062] In step five, the audio-visual prompts are: a green light for low risk and a prompt tone at a frequency of 1 time per minute; a yellow light for medium risk and a prompt tone at a frequency of 2 times per minute; and a red light for high risk and a prompt tone at a frequency of 3 times per minute.

[0063] By clearly identifying the audio-visual prompts corresponding to low, medium, and high risk levels, construction workers can quickly and intuitively identify the risk level, avoiding reaction delays caused by ambiguous warning signals, improving the timeliness and recognizability of risk warnings, and facilitating construction workers to quickly initiate corresponding response measures.

[0064] In step three, the water level sensor and stress sensor must be waterproofed before installation, with a waterproof rating of IP68. The sensor connection cable must be armored to avoid data transmission interruption due to damage during construction. When laying the cable, it must be fixed inside the tunnel segment with a fixing interval of ≤1 meter.

[0065] By protecting the sensors and connecting cables, data transmission interruptions can be avoided due to equipment failure or damage to the sensor connecting cables, ensuring that data collection is not interrupted during the dynamic risk assessment process.

[0066] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0067] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A dynamic risk assessment method for karst caves in large-diameter shield tunnels, characterized in that, The specific steps are as follows: Step 1: Pre-construction basic data collection and assessment model building: Collect karst geology, tunnel design and construction parameters, and build an initial risk assessment model containing core indicators and initial weights; Step 2: Initial risk level determination before construction: Substitute the data and calculate the initial risk value using the weighted summation method, classify it into low / medium / high risk levels, and formulate corresponding initial response preparation measures; Step 3: Real-time data acquisition during construction: Collect real-time data on shield machine operating parameters, karst cave detection data, water level, ground settlement, and segment stress at the prescribed frequency; Step 4: Real-time data processing and evaluation indicator update: Preprocess real-time data, update core evaluation indicators, and dynamically adjust indicator weights; Step 5: Dynamic Risk Value Calculation and Level Determination: Substitute the updated indicators and weights, calculate the real-time risk value every minute, and lock the level and provide an audio-visual alert after 5 consecutive minutes of stability; Step Six: Dynamic Adjustment and Implementation of Risk Response Measures: Implement corresponding response measures such as shield machine parameter control, monitoring frequency adjustment, and grouting filling according to low / medium / high risk levels; Step 7: Data Recording and Evaluation Model Optimization: Record data daily to form a daily report archive. After each 100-meter tunnel construction is completed, summarize the data, calibrate the weights, supplement the quantitative standards, and optimize the evaluation model.

2. The dynamic risk assessment method for karst caves in large-diameter shield tunnels according to claim 1, characterized in that, The specific steps of step one are as follows: In accordance with tunnel construction design specifications, karst geological data were collected within a 50-meter radius of the tunnel axis, including the distribution range of soluble rocks, rock layer thickness, rock layer dip angle, groundwater level, groundwater flow velocity, and groundwater flow direction. Drilling exploration was conducted using 150 mm diameter drilling equipment, with one exploration borehole set up every 20 meters to obtain data on karst cave development within the boreholes, including the depth of cave occurrence, transverse diameter of the cave, longitudinal length of the cave, type of cave filling material, and density of the filling material, calculated by weight, with values ​​ranging from 0-100%. Collect design and construction parameters for large-diameter shield tunnels, including tunnel inner diameter, tunnel outer diameter, tunnel burial depth, shield machine model, shield machine design parameters, and segment design parameters; An initial risk assessment model was established, and core assessment indicators and initial weights were determined. Core indicators included the minimum distance between the karst cave and the tunnel structure, the volume of the karst cave, and the stability of the karst cave filling. For mud-filled karst caves, the water content of the silt was classified as follows: ≤30% was stable, 30%-60% was relatively stable, and >60% was unstable. For water-filled karst caves, the difference in water level between the cave and the tunnel was classified as follows: ≤5 meters was stable, 5-10 meters was relatively stable, and >10 meters was unstable. Empty karst caves were directly classified as unstable. Ground subsidence warning values ​​were set at ≤3 mm for residential areas and ≤5 mm for municipal road areas. The deviation values ​​of the tunnel boring machine's propulsion parameters were calculated using data from over 100 accident cases and normal construction data of large-diameter shield tunnel projects in karst areas. The initial weights of each indicator were calculated using a data fitting method, with the following weights: minimum distance between karst cave and tunnel (0.25), karst cave volume (0.2), infill stability (0.2), ground settlement warning value (0.15), and tunnel boring machine propulsion parameter deviation value (0.2).

3. The dynamic risk assessment method for karst caves in large-diameter shield tunnels according to claim 1, characterized in that, The specific steps for step two are as follows: Substitute the basic geological data and tunnel parameters collected in Step 1 into the initial risk assessment model, and calculate the initial risk value using a weighted summation method: First, convert each indicator into a quantitative score according to the "risk level - numerical value" correspondence: minimum distance between the karst cave and the tunnel <2 meters is 100 points, 2-5 meters is 70 points, 5-10 meters is 40 points, and >10 meters is 10 points; karst cave volume >100 cubic meters is 100 points, 50-100 cubic meters is 70 points, 10- 50 cubic meters = 40 points, <10 cubic meters = 10 points; Unstable infill material = 100 points, relatively stable = 50 points, stable = 10 points; Ground settlement warning value: Actual monitored value ≥ warning value = 100 points, 0.5 × warning value - warning value = 50 points, <0.5 × warning value = 10 points; Tunnel boring machine propulsion parameter deviation >20% = 100 points, 10%-20% = 50 points, <10% = 10 points. Then, the scores of each indicator are multiplied by their corresponding initial weights and summed to obtain the initial risk value, which ranges from 0 to 100. The initial risk levels are divided as follows: 0-30 points is low risk level, 31-60 points is medium risk level, and 61-100 points is high risk level. Initial response preparation measures are formulated for different risk levels. For low-risk levels, settlement monitoring instruments and water level monitoring instruments are prepared. For medium-risk levels, ground-penetrating radar is added to the low-risk equipment. For high-risk levels, grouting pumps, grouting materials and quick-setting concrete are prepared in addition to the medium-risk equipment.

4. The dynamic risk assessment method for karst caves in large-diameter shield tunnels according to claim 1, characterized in that, Step three is as follows: The tunnel boring machine (TBM) utilizes its built-in sensor system to collect operating parameters every minute, including actual thrust, torque, propulsion speed, excavated soil volume, cutterhead rotation speed, and TBM attitude parameters. This data is transmitted in real-time to the tunnel monitoring center database. A 250 MHz ground-penetrating radar is used to conduct supplementary karst cave detection in front of the TBM cutterhead and on both sides of the tunnel. Detection is performed every 5 meters of tunnel construction length, acquiring the real-time location, size, and filling status of the caves. The detection data is then processed by the radar. After being processed by the software, the data is transmitted to the monitoring center database every 5 minutes. A water level monitoring point is set up every 10 meters along the tunnel axis, and a water level sensor is used to collect the water level height in the tunnel every 30 minutes. A settlement monitoring point is set up every 20 meters along the ground axis above the tunnel, and a high-precision settlement monitoring instrument is used to collect the ground settlement every hour. A stress monitoring point is set up every 5 rings at the tunnel segment circumferential joints, and a stress sensor is used to collect the segment stress value every 2 hours. All monitoring data is uploaded to the monitoring center database in real time.

5. The dynamic risk assessment method for karst caves in large-diameter shield tunnels according to claim 1, characterized in that, The specific steps for step four are as follows: The real-time data collected in step three is preprocessed. Values ​​exceeding the mean ± 3 standard deviations are identified as outliers. Outliers caused by equipment malfunctions are removed. Then, a moving average method with a window size of 5 collection periods is used to smooth the data after outlier removal, resulting in a stable real-time data sequence. Based on the processed real-time data, the core evaluation indicators in step one are updated: the minimum distance between the karst cave and the tunnel structure is recalculated according to the real-time location detected by ground-penetrating radar; the volume of the karst cave is recalculated according to the real-time dimensions; the stability of the karst cave filling is reassessed according to the real-time status; and the ground subsidence warning value is updated according to the ratio of the real-time ground subsidence to the warning value. The deviation values ​​of the tunnel boring machine's propulsion parameters are updated based on the real-time difference between the actual parameters and the rated parameters. The weights of the indicators are adjusted according to the risk correlation reflected by the real-time data: when the ground settlement reaches 80% or more of the warning value, the weight of the ground settlement warning value is adjusted from 0.15 to 0.3; when the deviation value of the tunnel boring machine's propulsion parameters reaches 15% or more of the rated value, the weight of the propulsion parameter deviation value is adjusted from 0.2 to 0.3; when the minimum distance between the karst cave and the tunnel is <3 meters, the weight of this indicator is adjusted from 0.25 to 0.35; in other cases, the initial weights are maintained, and the weight adjustments ensure that the sum of the weights of all indicators is 1.

6. The dynamic risk assessment method for karst caves in large-diameter shield tunnels according to claim 1, characterized in that, The specific steps for step five are as follows: The updated assessment indicators from step four are quantified into scores, using the same quantification standards as in step two. These scores are then substituted into the assessment model along with the dynamically adjusted weights. A weighted summation method is used to calculate the real-time risk value, with the calculation frequency matching the tunnel boring machine parameter acquisition frequency, maintaining a frequency of once per minute. The real-time risk value is simultaneously displayed on the monitoring center console. The current risk level is determined according to the level classification standards of step two. When the real-time risk value remains within a certain level range for 5 consecutive minutes, the system automatically locks that level and issues an audible and visual alert. If the real-time risk value fluctuates across levels within 5 minutes, the system does not lock the level temporarily, but continues to calculate and display the fluctuation trend until it stabilizes at a certain level for 5 consecutive minutes.

7. The dynamic risk assessment method for karst caves in large-diameter shield tunnels according to claim 1, characterized in that, The specific steps for step six are as follows: When the system locks in a low-risk level, maintain the tunnel boring machine's normal advance parameters: advance speed 20-30 mm / min, cutterhead speed 2-3 rpm, and excavation volume controlled within ±5% of the design value. Continue collecting various data at the original frequency, summarizing and analyzing the monitoring data daily. After confirming no abnormal changes in risk indicators, proceed with construction as planned the following day. When the system locks in a medium-risk level, immediately reduce the tunnel boring machine's advance speed to 10-15 mm / min, cutterhead speed to 1-2 rpm, and control the excavation volume to 80%-90% of the design value. Increase the frequency of ground-penetrating radar detection to once every 2 meters and the frequency of ground settlement monitoring to once every 30 minutes. Assign dedicated personnel to manually check the tunnel boring machine's attitude and segment stress data every hour. If the segment stress value is found to exceed 90% of the design value, immediately... The tunnel boring machine (TBM) should be suspended and the cause investigated. Progress can only resume after the investigation is complete and the data returns to normal. If the system locks in a high-risk level, the TBM should be stopped immediately, the cutterhead drive system shut down, and emergency drainage equipment with a drainage capacity of ≥50 cubic meters per hour activated. The grouting pump should be started, and the grouting volume calculated as 1.2 times the volume of the karst cave should be used. Cement grout should be injected into the karst cave through the grouting holes on the TBM cutterhead. Pressure data inside the karst cave should be collected every 10 minutes during grouting. Grouting should be stopped when the pressure stabilizes at 0.3-0.5 MPa. After grouting, the tunnel should be left to stand for 24 hours. Then, the filling condition of the karst cave should be detected using ground-penetrating radar. Once the filling density is confirmed to be ≥90%, the real-time risk value should be recalculated. If the risk value drops below 30 points, progress should be resumed according to low-risk level parameters, with an initial progress speed of 5 mm / min, gradually increasing to the normal speed. If the risk value is still higher than 30, repeat the grouting process until the risk level drops to low risk.

8. The dynamic risk assessment method for karst caves in large-diameter shield tunnels according to claim 1, characterized in that, Step 7 involves the following steps: After each day's construction, detailed records are kept of the day's basic data, real-time collected data, data processing results, risk value calculation process, risk level determination results, and the implementation and effectiveness of countermeasures, i.e., changes in risk values ​​and settlement before and after grouting. This forms a daily dynamic risk assessment report, which includes data tables, risk level change curves, photos of countermeasure implementation, and text descriptions. All daily reports are archived and stored according to the construction date. After every 100 meters of tunnel construction is completed, all daily report data for that section are summarized, and regression analysis is used to recalibrate the weights of the assessment indicators. Indicators with weight deviations exceeding 5% are adjusted. Based on the type of karst cave encountered in that section, specific quantitative standards for that type of karst cave are supplemented. The optimized model is used for dynamic risk assessment of subsequent sections to ensure that the model is adapted to the karst geological characteristics of different construction sections.

9. The method for dynamic risk assessment of karst caves in large-diameter shield tunnels according to claim 1, characterized in that: The drilling equipment mentioned in step one must be a hydraulic core drilling rig. The verticality deviation of the drilling rig should be controlled within ≤0.5%. Core samples should be taken every 5 meters during the drilling process, and the core recovery rate should be ≥90%. In the preferred step one, the accident case data and normal construction data of the large-diameter shield tunnel project in the karst area should be selected from large-diameter shield tunnel projects in the karst area completed within the past 10 years with a geological condition similarity of ≥80%, to ensure the reference value of the data. Preferably, the shield machine sensors mentioned in step three need to be calibrated once a month, and the calibration error is controlled within ±2%; before the ground-penetrating radar detection, it needs to be debugged in a simulated karst cave environment of known size to ensure that the detection error is ≤5%, and the distance between the radar antenna and the tunnel wall remains stable during detection, with a deviation of ≤2 cm; Preferably, the water level sensor and stress sensor mentioned in step three need to be waterproofed before installation, with a waterproof rating of IP68. The sensor connection cable should be armored cable to avoid data transmission interruption due to damage during construction. When laying the cable, it should be fixed inside the tunnel segment with a fixing spacing of ≤1 meter. Preferably, when applying the processed real-time data described in step four, the collected data must first undergo a normality test, and the criterion for removing outliers can only be used if the pass rate of the test is ≥95%. If your choice is correct, the weight adjustment in step four needs to be verified using more than 5 sets of historical risk case data. If the verification accuracy is ≥90%, the adjustment is confirmed to be effective; otherwise, the weight calculation method needs to be re-optimized. In the preferred step five, the audio-visual prompts are: a green light for low risk and a prompt tone at a frequency of 1 time per minute; a yellow light for medium risk and a prompt tone at a frequency of 2 times per minute; and a red light for high risk and a prompt tone at a frequency of 3 times per minute. Preferably, the cement used for the grouting material in step six should be 42.5 grade ordinary Portland cement, and the admixture should be a high-efficiency water-reducing agent; the geological radar detection after grouting should cover the entire area of ​​the karst cave, with the detection point spacing ≤ 1 meter, to ensure comprehensive verification of the filling density.

Images