A roadway roof anchoring state intelligent evaluation method

Abstract

A kind of roadway roof anchoring state intelligent evaluation method belongs to coal mine roadway bolting monitoring technical field, lays out joint monitoring section, synchronously collects roof separation displacement and anchor rod axial force data and transmits to ground server;After data pre-processing, calculate separation rate and anchor rod axial force increment rate according to self-adaptive sliding window;According to the influence range of anchor rod, three sections are divided, and three kinds of stress models containing instantaneous elasticity and viscous lag response are established;Determine the separation position by goodness of fit comparison and output the optimal parameters;Construct dimensionless support efficiency comprehensive index SEI;Based on historical normal data, calculate dynamic threshold;Combined with separation position, SEI and threshold comparison and parameter trend, divide seven risk levels;Output corresponding early warning and disposal suggestion and push in real time.The method is simple in implementation process and low in implementation cost, which can intelligently identify the real position of separation and construct support efficiency comprehensive index and dynamic grading early warning.

Description

Technical Field

[0001] This invention belongs to the field of coal mine roadway anchor support monitoring technology, specifically relating to an intelligent evaluation method for the anchorage status of roadway roof. Background Technology

[0002] As the depth of mining in my country continues to extend and the intensity of mining continues to increase, the environment of the surrounding rock in roadways is becoming increasingly complex. Extreme characteristics such as high ground stress, strong mining disturbance, and soft rock rheology are becoming increasingly prominent, which puts forward stringent requirements on the quality of roadway support and roof stability.

[0003] With its outstanding advantages such as active load-bearing capacity, convenient construction, economic efficiency, and strong adaptability, rock bolt support has become a core technical means for roof control in roadways and is widely used in various mine roadway support projects. In the monitoring and evaluation of rock bolt support effectiveness, roof delamination and rock bolt stress are two of the most direct and crucial monitoring indicators. Roof delamination directly reflects the degree of deformation and evolution trend of the surrounding rock, while rock bolt stress accurately characterizes the actual load-bearing state of the support structure. Together, they constitute the core basis for evaluating the effectiveness of rock bolt support and judging roof stability.

[0004] In recent years, with the rapid development and popularization of sensor technology and mine Internet of Things (IoT) technology, monitoring equipment such as roof delamination meters and anchor bolt force gauges have been widely used in many mines across China. These devices have collected and accumulated a large amount of real-time monitoring data on roof delamination and anchor bolt stress, laying a data foundation for dynamic monitoring, intelligent sensing, and scientific evaluation of roadway roof conditions. They have also provided support for optimizing support effectiveness and predicting safety risks. However, current research and engineering applications of roadway roof monitoring technology still have significant shortcomings, making it difficult to meet the actual needs of accurate evaluation of roof anchorage status and safety risk prevention under complex surrounding rock conditions. Specifically, existing monitoring systems generally deploy roof delamination meters and anchor bolt force gauges as independent monitoring units. The data acquisition, storage, analysis, and alarm functions of the two types of equipment are basically independent, lacking deep integration and forming data silos, thus hindering the collaborative analysis and integrated application of monitoring data.

[0005] In practical field applications, when both roof delamination exceeding limits and anchor bolt stress exceeding limits occur simultaneously, field technicians often struggle to accurately quantify the inherent causal relationship between the two. They cannot precisely determine whether anchor bolt failure leads to the continued development of roof delamination, or whether the exacerbation of roof delamination causes abnormally increased anchor bolt stress. Furthermore, they cannot reliably distinguish whether the anchor bolt is under effective load-bearing capacity but the roof delamination occurs outside the anchoring zone, or whether a hidden problem in the anchoring system itself leads to both exceeding limits. This ambiguity in causality makes it difficult to translate monitoring alarm information into effective engineering decision-making, resulting not only in widespread false alarms and missed alarms, but also in reduced trust in alarm information among field workers, leading to delayed responses, neglect of potential safety risks, and ultimately creating significant hidden dangers for mine safety. Moreover, the specific location of roof delamination (e.g., within the free section of the anchor bolt, within the anchoring section, or outside the anchoring section) significantly affects roof stability, which is also a key factor in judging the actual working state of the anchor bolt and evaluating its anchoring effect. However, conventional single-point / double-point roof delamination meters widely used in the field can only monitor the total amount of delamination, but cannot effectively identify the specific location where delamination occurs. This makes it impossible to accurately evaluate the actual working state of the anchor bolts and the effectiveness of the anchoring system, making it difficult to formulate scientific and accurate support optimization and risk control measures. This further restricts the application effect of anchor bolt support technology under complex surrounding rock conditions.

[0006] In summary, given the shortcomings of current roadway roof monitoring, such as data silos, ambiguous causal relationships, and inaccurate identification of delamination locations, there is an urgent need to propose a scientific and efficient intelligent evaluation method for the anchorage status of roadway roofs. This method would enable the integrated application of monitoring data, accurate identification of delamination evolution characteristics, and scientific evaluation of anchorage status, thus providing reliable technical support for mine roof safety control. Summary of the Invention

[0007] To address the problems existing in the prior art, this invention provides an intelligent evaluation method for the anchorage status of roadway roof. This method is simple to implement and has low implementation costs. It can intelligently identify the true location of the delamination and construct a comprehensive index of support effectiveness and dynamic hierarchical early warning.

[0008] To achieve the above objectives, the present invention provides an intelligent evaluation method for the anchorage status of roadway roof, comprising the following steps:

[0009] Step 1: Joint monitoring section layout and data acquisition; roof delamination meter and anchor bolt force gauge are simultaneously deployed in the roof monitoring section of the same roadway to collect roof delamination displacement and anchor bolt axial force data in real time, and transmit them to the ground server through the mine ring network;

[0010] Step 2: Data preprocessing and feature calculation; the raw data is filtered, denoised, and imputed, and then the separation rate and anchor bolt axial force increment rate are calculated according to the sliding time window, and the window length and reset strategy are adaptively adjusted according to the tunneling disturbance.

[0011] Step 3: Constructing multiple anchor bolt force models based on differences in mechanical response; According to the roof surrounding rock structure and anchor bolt anchoring characteristics, the influence range of the anchor bolt is divided into the free section, the anchoring section, and the section outside the anchor bolt. An anchor bolt force models are established for the three assumed delamination positions. Each model is composed of the superposition of instantaneous elastic response and viscous hysteresis response.

[0012] Step 4: Detached location determination based on goodness-of-fit comparison; within the same sliding window, extract measured data at several time points, perform nonlinear least squares fitting using three models respectively, take the detached location corresponding to the model with the smallest residual as the current determination result, and output the optimal model parameters;

[0013] Step 5: Construction of the comprehensive support effectiveness index; Based on the optimal model parameters and fitting residuals obtained from the inversion, a dimensionless comprehensive support effectiveness index (SEI) is constructed to quantitatively evaluate the current working status of the anchoring system.

[0014] Step 6: Dynamic threshold calibration; Based on the SEI sequence of historical normal support stages, a moving average method is used to calculate the dynamic threshold, which is used to adaptively determine whether the current support status has deteriorated;

[0015] Step 7: Determine the support status; Based on the identified delamination location, the comparison between the current SEI and the dynamic threshold, and the changing trend of the optimal model parameters, the support effect is divided into seven risk levels from low to high.

[0016] Step 8: Output of graded early warning and response suggestions; based on the determined status, output the corresponding early warning level and response suggestions, and push them to relevant personnel and equipment in real time through various communication methods.

[0017] As a preferred option, the process of joint monitoring section layout and data acquisition in step 1 is as follows:

[0018] S11: Deployment of roof separation instrument and anchor bolt force gauge; In the same roadway roof monitoring section, the roof separation instrument is deployed in the roof separation instrument borehole in the middle of the roof, and the anchor bolt force gauge is installed on the adjacent anchor bolt trays on both sides of the separation instrument borehole. The horizontal distance between the adjacent anchor bolts and the separation instrument borehole is less than 0.5m to ensure that both reflect the surrounding rock deformation and anchor bolt force at the same location;

[0019] S12: Synchronous data acquisition; Data is acquired synchronously using a roof separation instrument and a bolt force gauge, and the data is first stored in the downhole monitoring substation;

[0020] S13: Data transmission; Data from underground monitoring substations is transmitted to the ground server in real time via the mine ring network;

[0021] S14: Output variable; Top plate delamination meter outputs the relative delamination displacement of the shallow base point relative to the deep base point. The anchor bolt force gauge outputs the axial force of the anchor bolt. .

[0022] As a preferred embodiment, the data preprocessing and feature calculation process in step 2 is as follows:

[0023] S21: Preprocessing; for relative delamination displacement and anchor bolt axial force The original sequence is subjected to low-pass filtering for noise reduction and linear interpolation for missing values;

[0024] S22: Window length; defines the sliding time window length. The frequency of tunneling disturbance is selected within the range of 5 to 15 minutes. Specifically, 5 minutes is used when the tunneling cycle advance is large or the surrounding rock deformation rate is high, and 15 minutes is used in stable sections.

[0025] S23: Calculation rate; calculates the delamination rate within each window. and the rate of increase of anchor bolt axial force ;

[0026] S24: Disturbance Reset; When strong disturbances such as the tunneling machine cutting hard rock occur, the sliding window is automatically reset to eliminate sudden interference.

[0027] As a preferred option, in step 3, the process of constructing the multiple models of anchor bolt force based on the differences in mechanical response is as follows:

[0028] S31: Division of surrounding rock sections; Based on the structural characteristics of the anchor bolts, the surrounding rock of the roof is divided into three sections along the depth direction: the free section, the anchored section, and the section outside the anchor bolts.

[0029] S32: Establish anchor bolt force models; for the three sections where delamination may occur, establish anchor bolt force models under the assumed delamination locations:

[0030] Model A is constructed based on the following formula: ;

[0031] Model B is constructed based on the following formula: ;

[0032] Model C is constructed based on the following formula: ;

[0033] In the formula, , where is the delamination rate; The elastic modulus of the anchor bolt; This represents the cross-sectional area of ​​the anchor bolt. This refers to the length of the free section of the anchor bolt. This is the vertical distance from the delamination zone to the anchoring end of the anchor bolt; This is the total length of the anchor bolt; The equivalent interaction coefficient of the free segment; Shear stiffness at the anchorage section interface; This refers to the equivalent force transmission stiffness between the end of the anchor bolt and the deep delamination zone. is the rheological attenuation coefficient of the surrounding rock.

[0034] As a preferred embodiment, the delamination location determination process based on goodness-of-fit comparison in step 4 is as follows:

[0035] For each sliding window, take all consecutive values ​​within the window. The measured data at each moment ,in Substitute the data into models A, B, and C respectively, and use the nonlinear least squares method to invert the unknown parameter sets of each model. Model A Model B Model C The objective function is the sum of squared residuals, as shown in the following equation:

[0036] ;

[0037] In the formula, ; The model predicts the anchor bolt axial force; the minimum residuals of the three models are compared. , , The delamination location represented by the model with the smallest residual is taken as the discrimination result of the current window: if The minimum delamination occurs in the free segment; if The minimum occurs in the anchorage section; if The minimum occurs in the section outside the anchor bolt; at the same time, the parameter estimates under the optimal model are recorded. , or And calculate the standard deviation of the fitted residuals. ,in This represents the number of parameters to be estimated in the model.

[0038] As a preferred option, in step 4, the nonlinear least squares method employs the Levenberg-Marquardt algorithm, with the iteration termination condition being that the parameter change is less than 10. -6 Or the residual change is less than 10 -8 .

[0039] As a preferred option, the process of constructing the comprehensive support effectiveness index in step 5 is as follows:

[0040] For cases where the delamination is located within the free section or anchorage section, the comprehensive index of support effectiveness should be constructed according to the following formula. :

[0041] ;

[0042] In the formula, Pick or ; Fit the residual standard deviation to the corresponding model; The standard deviation of the measured values ​​of the anchor bolt axial force within the window; Used as a reference for shear stiffness in design;

[0043] For sections where the delamination is located outside the anchor bolts, the comprehensive index of support effectiveness should be constructed according to the following formula. :

[0044] ;

[0045] In the formula, This serves as a reference force transmission coefficient for design.

[0046] As a preferred embodiment, the dynamic threshold calibration process in step 6 is as follows:

[0047] Collect the current moment before A sliding window Historical values, calculate their mean and standard deviation The dynamic threshold is defined according to the following formula. :

[0048] ;

[0049] History window length Adaptive adjustments based on the tunnel service stage: [During excavation] During the recovery period After each strong disturbance, the statistical parameters are recalculated; when the current window... At that time, the support effectiveness was determined to be significantly lower than the historical normal level.

[0050] As a preferred embodiment, the support status determination process in step 7 is as follows:

[0051] Based on the delamination location determined in step 4, the current... With dynamic threshold Based on the magnitude relationship and the changing trend of the optimal model parameters, the following seven states are defined:

[0052] State I: The delamination is located within the free segment. ,and Stablize;

[0053] State II: Delamination is located within the free segment. ,and Continued decline;

[0054] State III: Delamination is located within the anchorage section. ,and Stablize;

[0055] State IV: Delamination is located within the anchorage section. ,and Continued decline;

[0056] State V: Delamination is located outside the anchor bolt. ,and Stablize;

[0057] State VI: Delamination is located outside the anchor bolt. ,and Continued decline;

[0058] State VII: Any position outside the layer. And model residuals .

[0059] As a preferred option, the output process for the graded early warning and response recommendations in step 8 is as follows:

[0060] Based on the status determined in step 7, the following warning level and handling recommendations are output:

[0061] When in State I, a Level I warning will be issued, along with a reminder to maintain normal monitoring.

[0062] When the status is II, a Level II warning will be issued, and a reminder message will be sent to strengthen observation and, if necessary, to install additional anchor bolts;

[0063] When the status is III, a Level II early warning will be issued, and a reminder message will be sent to increase the intensity of monitoring and pay attention to changes in the anchorage section;

[0064] When the status is IV, a Level III warning will be issued, along with a reminder message to reinforce the anchorage section with grouting or to install additional long anchor cables;

[0065] When the status is V, a Level III warning will be issued, along with a reminder to strengthen deep support and consider installing additional anchor cables;

[0066] When the status is VI, a Level IV warning will be issued, along with a reminder to immediately take deep reinforcement measures and prepare for a production shutdown.

[0067] When status VII is reached, a Level V warning will be issued, along with a reminder to immediately stop work, evacuate personnel, and fully reinforce the structure.

[0068] Meanwhile, the early warning information is pushed to the wellhead dispatch center and the mobile terminals of on-site personnel in real time, and the audible and visual alarms set up in key locations underground are used to carry out early warning actions.

[0069] Compared with the prior art, the present invention has the following technical advantages:

[0070] 1. A spatiotemporal synchronous joint monitoring system for delamination and stress was established, laying a reliable data foundation. By deploying the top plate delamination meter and the anchor bolt force gauge in the same borehole or adjacent boreholes with a spacing of less than 0.5m within the same monitoring section, synchronous acquisition and spatiotemporal correspondence of delamination displacement and anchor bolt axial force were achieved. This method effectively overcomes the shortcomings of traditional monitoring, which involves isolated acquisition of two types of key data and difficulty in correlation analysis, providing a highly reliable data foundation for subsequent delamination location identification and anchorage status evaluation.

[0071] 2. A multi-model inversion method based on differences in mechanical response was proposed, enabling intelligent identification of the delamination location. The surrounding rock of the roof was divided into a free section, an anchored section, and a section outside the anchor based on the structural characteristics of the anchor bolt. Physical models of the anchor bolt force (Models A, B, and C) were established for the three assumed delamination locations. Each model includes both instantaneous elastic response and viscous hysteresis response, accurately reflecting the attenuation effect of historical delamination rates on the current force. Nonlinear least-squares inversion was performed on each model using measured data within a sliding window. By comparing the fitting residuals, the actual location of the delamination (free section / anchored section / outside the anchor bolt) could be intelligently determined. This method fundamentally solves the technical problem of accurately locating delamination in traditional monitoring.

[0072] 3. A comprehensive support performance index considering the delamination location was constructed, enabling a quantitative and refined evaluation of the anchoring state. Based on the optimal model parameters (interface shear stiffness or equivalent force transmission coefficient) obtained through inversion and the fitting residuals, a dimensionless comprehensive support performance index (SEI) was constructed. This index integrates the delamination location, anchor bolt working stiffness, and model fitting reliability into the evaluation framework, quantitatively reflecting the degree to which the current anchoring system maintains its performance relative to the design level. Furthermore, a moving average method was used to calculate the dynamic threshold, achieving adaptive tracking of the threshold to changes in surrounding rock conditions.

[0073] 4. A seven-level risk status classification system was established, covering a complete working condition chain from normal to instability precursors. Combining three dimensions—delamination location, comparison of SEI with dynamic thresholds, and the trend of optimal model parameter changes (stable / continuously decreasing)—seven risk levels, from State I to State VII, were defined, ranging from low to high. This system not only covers typical working conditions where delamination occurs in the free section, anchored section, and anchor bolts, but also specifically sets up State VII (critical state), which can identify the overall instability precursor—a sudden change in physical relationships—when the model residual is abnormally large. Each state has clear mechanical significance and engineering implications, avoiding frequent false alarms caused by single threshold alarms.

[0074] 5. Effectively solves the core technical problems of frequent false alarms and insufficient decision-making basis in traditional monitoring. In traditional methods, delamination data and anchor bolt stress data are often analyzed separately, making it difficult to establish a causal relationship. This leads to frequent false alarms in situations where there is delamination change but the anchor bolt is not lost, or the anchor bolt stress increases but the delamination is stable. This invention, through the organic integration of intelligent delamination location identification and comprehensive evaluation of anchoring effectiveness, can accurately distinguish whether the delamination occurs in the free section, the anchored section, or outside the anchor bolt, corresponding to different physical processes such as shallow loosening, anchoring interface damage, and deep rock mass weakening. This provides clear and quantifiable basis for decisions on whether to install additional anchor bolts, grouting reinforcement, or additional anchor cables on site.

[0075] This method is simple to implement and has low implementation costs. It can intelligently identify the true location of the delamination and construct a comprehensive index of support effectiveness and dynamic hierarchical early warning, effectively solving the problems of isolated data, unclear causal relationships and frequent false alarms in traditional monitoring. Attached Figure Description

[0076] Figure 1 This is a flowchart from the present invention. Detailed Implementation

[0077] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0078] like Figure 1 As shown, this invention provides an intelligent evaluation method for the anchorage status of roadway roof, comprising the following steps:

[0079] Step 1: Joint monitoring section layout and data acquisition;

[0080] In the same roadway roof monitoring section, roof delamination instrument and anchor bolt force gauge are simultaneously deployed to collect roof delamination displacement and anchor bolt axial force data in real time, and transmit them to the ground server through the mine ring network.

[0081] As a preferred option, the process of joint monitoring section layout and data acquisition is as follows:

[0082] S11: Deployment of roof separation instrument and anchor bolt force gauge; In the same roadway roof monitoring section, the roof separation instrument is deployed in the roof separation instrument borehole in the middle of the roof, and the anchor bolt force gauge is installed on the adjacent anchor bolt trays on both sides of the separation instrument borehole. The horizontal distance between the adjacent anchor bolts and the separation instrument borehole is less than 0.5m to ensure that both reflect the surrounding rock deformation and anchor bolt force at the same location;

[0083] S12: Synchronous data acquisition; Data is acquired synchronously using a roof separation instrument and a bolt force gauge. The sampling frequency can be no less than 1Hz. The data is first stored in the downhole monitoring substation.

[0084] S13: Data transmission; Data from underground monitoring substations is transmitted to the ground server in real time via the mine ring network;

[0085] S14: Output variable; Top plate delamination meter outputs the relative delamination displacement of the shallow base point relative to the deep base point. (Unit: mm) Anchor bolt force gauge outputs anchor bolt axial force. (Unit: kN).

[0086] In this technical solution, by placing the roof delamination meter and the anchor bolt force gauge in the same borehole or adjacent boreholes with a very small spacing, the spatial positional correspondence and high-frequency synchronous acquisition of the delamination displacement and anchor bolt axial force are realized. The data is uploaded in real time with the help of underground substations and mine ring networks. Without increasing the investment in dedicated hardware, this solution effectively solves the problem of isolation and spatiotemporal misalignment of the two types of key data in traditional monitoring, laying a reliable and unified data foundation for subsequent intelligent identification of delamination location and dynamic evaluation of anchoring status.

[0087] Step 2: Data preprocessing and feature calculation;

[0088] The raw data is filtered, denoised, and imputed. Then, the separation rate and anchor bolt axial force increment rate are calculated according to the sliding time window. The window length and reset strategy are adaptively adjusted according to the tunneling disturbance.

[0089] As a preferred approach, the data preprocessing and feature calculation process is as follows:

[0090] S21: Preprocessing; for relative delamination displacement and anchor bolt axial force The original sequence is subjected to low-pass filtering for noise reduction (such as moving average or median filtering) and linear imputation for missing values;

[0091] S22: Window length; defines the sliding time window length. The frequency of tunneling disturbance is selected within the range of 5 to 15 minutes. Specifically, 5 minutes is used when the tunneling cycle advance is large or the surrounding rock deformation rate is high, and 15 minutes is used in stable sections to eliminate sudden disturbances.

[0092] S23: Calculation rate; calculates the delamination rate within each window. (Unit: mm / min) and anchor bolt axial force increment rate (Unit: kN / min);

[0093] S24: Disturbance Reset; When strong disturbances such as the tunneling machine cutting hard rock occur, the sliding window is automatically reset (the current window's historical data is cleared, and the new window is recalculated with the end of the disturbance as the starting point) to eliminate sudden disturbances.

[0094] In this technical solution, the quality and continuity of delamination displacement and anchor bolt axial force data are ensured by performing low-pass filtering and linear interpolation on the original sequence. A sliding time window that can be adaptively adjusted within 5 to 15 minutes according to the frequency of tunneling disturbances is adopted, and an automatic window reset mechanism is set for strong disturbances such as the tunneling machine cutting hard rock. This can effectively eliminate the influence of sudden interference on feature quantity calculation, thereby stably and accurately extracting the delamination rate and anchor bolt axial force increment rate, providing reliable dynamic feature input for subsequent multi-model inversion and intelligent identification of delamination location.

[0095] Step 3: Constructing multiple models of anchor bolt force based on differences in mechanical response;

[0096] Based on the structure of the surrounding rock of the roof and the anchoring characteristics of the anchor, the influence range of the anchor is divided into the free section, the anchoring section and the section outside the anchor. An anchor force model is established for the three assumed delamination positions. Each model is composed of the superposition of instantaneous elastic response and viscous hysteresis response.

[0097] As a preferred approach, the process of constructing multiple models of anchor bolt force based on differences in mechanical response is as follows:

[0098] S31: Surrounding Rock Section Division; Based on the characteristics of the anchor bolt structure, the surrounding rock of the roof is divided into three sections along the depth direction, namely the free section (depth). (No bond between the anchor bolt and the surrounding rock) Anchorage section (depth) (resin bonding), sections outside the anchor bolt (depth) (surrounding rock self-supporting zone);

[0099] S32: Establish anchor bolt force model; for the three sections where delamination may occur, establish anchor bolt force model under the assumed delamination position: each model is composed of instantaneous elastic response (proportional to the current delamination displacement D(t)) and viscous hysteresis response (decayed convolution with the historical delamination rate);

[0100] Model A (with delamination in the free segment) is constructed based on the following formula: ;

[0101] Model B (delamination in the anchorage section) is constructed based on the following formula: ;

[0102] Model C (for sections where delamination occurs outside the anchor bolts) is constructed based on the following formula: ;

[0103] In the formula, , is the delamination rate, in mm / s or mm / min, and must be consistent with the time unit; This refers to the elastic modulus of the anchor bolt, expressed in MPa or GPa. This represents the cross-sectional area of ​​the anchor bolt, in mm². The unit is N; This represents the length of the free section of the anchor bolt, in meters (m). This is the vertical distance from the separation zone to the anchorage end of the anchor bolt, in meters (m). This refers to the total length of the anchor bolt, in meters (m). The equivalent interaction coefficient for the free segment, expressed in N / mm, represents the increment of anchor bolt axial force caused by a unit delamination displacement in the free segment; The shear stiffness of the anchorage section interface is expressed in N / mm, representing the interface shear force caused by a unit relative displacement of the anchorage section. This is the equivalent force transmission stiffness between the end of the anchor bolt and the deep delamination zone, expressed in N / mm. The rheological attenuation coefficient of the surrounding rock is expressed in seconds. -1 This reflects the rate of decay of the influence of historical rates; among them, , , , Given constants, , , , The parameters to be inverted, Initial values ​​are assigned according to the surrounding rock type: 0.01–0.05s for soft rock. -1 For hard rock, the range is 0.001–0.01 s. -1 The exponent kernel in the integral term This indicates that the further away the delamination rate is from the current moment, the smaller its contribution to the current force.

[0104] In this technical solution, the surrounding rock of the roof is finely divided into a free section, an anchoring section, and a section outside the anchor bolt. For three assumed delamination locations, physical models of anchor bolt force (models A / B / C) are constructed, which are superimposed with instantaneous elastic response and viscous hysteresis response (including exponential decay convolution of historical delamination rate). This enables a quantitative characterization of the differences in the mechanical response of the anchor bolt under different delamination locations. The model introduces the surrounding rock rheological decay coefficient α, which can reflect the time cumulative effect of the creep characteristics of the surrounding rock on the anchor bolt force. The three established models constitute a computable comparison framework, providing a scientific theoretical basis and clear mathematical expression for subsequent intelligent identification of the actual location of delamination by comparing the goodness of fit.

[0105] Step 4: Determination of delamination location based on goodness-of-fit comparison;

[0106] Within the same sliding window, measured (D,F) data at several time points are extracted, and nonlinear least squares fitting is performed using three models respectively. The delamination position corresponding to the model with the smallest residual is taken as the current discrimination result, and the optimal model parameters are output.

[0107] As a preferred method, the delamination location determination process based on goodness-of-fit comparison is as follows:

[0108] For each sliding window, take all consecutive values ​​within the window. The measured data at each moment ,in (Corresponding to a window length of 5–15 minutes, sampling frequency ≥1Hz); Substitute the data into models A, B, and C respectively, and use the nonlinear least squares method to invert the unknown parameter set of each model. Model A Model B Model C The objective function is the sum of squared residuals, as shown in the following equation:

[0109] ;

[0110] In the formula, ; For the anchor bolt axial force predicted by the model, the integral term is approximated using numerical discretization (such as the trapezoidal rule or Euler method); as a preferred method, the nonlinear least squares method employs the Levenberg-Marquardt algorithm, with the iteration termination condition being that the parameter change is less than 10. -6 Or the residual change is less than 10 -8 Compare the minimum residuals of the three models. , , The delamination location represented by the model with the smallest residual is taken as the discrimination result of the current window: if The minimum delamination occurs in the free segment; if The minimum occurs in the anchorage section; if The minimum occurs in the section outside the anchor bolt; at the same time, the parameter estimates under the optimal model are recorded. , or And calculate the standard deviation of the fitted residuals. ,in This represents the number of parameters to be estimated in the model (p=2).

[0111] In this technical solution, N consecutive measured delamination displacement and anchor bolt axial force data pairs within a sliding window are used to perform nonlinear least squares inversion on models A / B / C respectively (using the Levenberg-Marquardt algorithm with strict iteration termination conditions). The actual location of the delamination is intelligently identified by directly comparing the minimum residual sum of squares of the three models, realizing the objectivity and automation of delamination location and completely eliminating the subjectivity of traditional methods that rely on experience. At the same time, the estimated values ​​of key parameters and the standard deviation of the fitting residuals under the optimal model are output, providing direct and accurate input for the subsequent quantitative calculation of the support effectiveness index (SEI) and model reliability assessment.

[0112] Step 5: Construction of comprehensive indicators for support effectiveness;

[0113] Based on the optimal model parameters and fitting residuals obtained from the inversion, a dimensionless support effectiveness comprehensive index SEI is constructed to quantitatively evaluate the current working status of the anchoring system.

[0114] As a preferred option, the process for constructing the comprehensive index of support effectiveness is as follows:

[0115] For cases where the delamination is located within the free section or anchorage section, the comprehensive index of support effectiveness should be constructed according to the following formula. :

[0116] ;

[0117] In the formula, Pick (Separation in the free segment) or (Delamination occurs in the anchorage section); Fit the residual standard deviation to the corresponding model; The standard deviation of the measured values ​​of the anchor bolt axial force within the window; The reference shear stiffness is determined by the average value of no less than three sets of field pull-out tests;

[0118] For sections where the delamination is located outside the anchor bolts, the comprehensive index of support effectiveness should be constructed according to the following formula. :

[0119] ;

[0120] In the formula, For design reference force transmission coefficient, empirical values ​​are taken according to the surrounding rock type: (0.8~1.0)×10 for intact rock mass. 8 N / m, for fractured rock mass, take (0.3~0.6)×10 8 N / m. A higher value indicates that the support performance is closer to the design level; when When the value inside the parentheses is zero or negative, then... This indicates that the model has failed or the support is severely abnormal.

[0121] In this technical solution, the corresponding inversion stiffness is applied according to the determined delamination location (free section / anchored section or outside the anchor bolt). or (or ) and design reference value ( or The ratio of ) is used as the core performance factor, while introducing ( As a model fitting reliability penalty term, a dimensionless comprehensive support effectiveness index (SEI) was constructed. This index can not only quantitatively evaluate the performance retention of the current anchoring system relative to the design level, but also... The system automatically identifies model mismatch or severe support anomalies (SEI≤0), enabling refined, normalized, and comparable evaluation of anchorage status. This provides a scientific and continuous quantitative basis for subsequent dynamic threshold calibration and graded early warning.

[0122] Step 6: Dynamic threshold calibration;

[0123] Based on the SEI sequence of historical normal support stages, a dynamic threshold is calculated using the moving average method to adaptively determine whether the current support status has deteriorated.

[0124] As a preferred option, the dynamic threshold calibration process is as follows:

[0125] Collect the current moment before A sliding window Historical values ​​(including only normal data before it was identified as abnormal or subject to human intervention) are used to calculate its mean. and standard deviation The dynamic threshold is defined according to the following formula. :

[0126] ;

[0127] History window length Adaptive adjustments based on the tunnel service stage: [During excavation] During the recovery period After each strong disturbance (such as a tunnel boring machine cutting through hard rock), the statistical parameters are recalculated (clearing the data before the disturbance and accumulating from the first window after the disturbance). (multiple windows) to eliminate the impact of abnormal events on the threshold; when the current window... At that time, the support effectiveness was determined to be significantly lower than the historical normal level.

[0128] In this technical solution, the mean and standard deviation of the historical SEI values ​​(including only normal data) of the M sliding windows prior to the current time are collected, and a dynamic threshold is defined. This method enables real-time adaptive tracking of the normal fluctuation range of support effectiveness; the historical window length M is automatically adjusted according to the roadway service stage (M=100 during tunneling, M=50 during mining), and the statistical parameters are reset after each strong disturbance, effectively eliminating the interference of tunneling disturbance or human intervention on threshold calculation; this method avoids the drawbacks of frequent false alarms or missed alarms due to changes in surrounding rock conditions caused by traditional fixed thresholds, and significantly improves the accuracy of anomaly detection and adaptability to complex working conditions.

[0129] Step 7: Determine the support status;

[0130] Based on the identified delamination location, the comparison between the current SEI and the dynamic threshold, and the changing trend of the optimal model parameters, the support effect is divided into seven risk levels from low to high.

[0131] As a preferred option, the support condition determination process is as follows:

[0132] Based on the delamination location determined in step 4, the current... With dynamic threshold The magnitude relationship, and the optimal model parameters ( , or The changing trend of the parameter (“stable” means that the parameter fluctuation does not exceed ±5% within three consecutive windows and there is no continuous unidirectional trend; “continuous decline” means that the parameter decreases monotonically within three consecutive windows and the cumulative decrease is greater than 10%) is defined in the following seven states:

[0133] Status I (Low Risk): The delamination is located within the free segment. ,and Stable indicates that the free section of the anchor bolt is working normally, the delamination is limited to the shallow part, and the top plate is stable;

[0134] State II (Lower Risk): Delamination is located within the free segment. ,and The continuous decline indicates that the performance of the free section anchor bolt is deteriorating, and closer monitoring is needed;

[0135] Status III (Medium Risk): Delamination is located within the anchorage section. ,and The stability indicates that the anchorage section can still effectively bear the load, but the delamination has extended to the anchorage area, and the development trend needs to be monitored.

[0136] Status IV (Higher Risk): Delamination occurs within the anchorage section. ,and A continuous decline indicates damage to the anchorage interface, a decrease in the anchor bolt's load-bearing capacity, and the need for timely reinforcement.

[0137] Status V (High Risk): Delamination occurs outside the anchor bolt. ,and Stable indicates that the deep rock mass is temporarily stable, but the delamination has exceeded the control range of the anchor bolt, posing a potential risk.

[0138] Condition VI (Severe Risk): Delamination occurs outside the anchor bolt. ,and The continued decline indicates that the deep rock mass has weakened and there is a possibility of overall collapse, requiring immediate deep reinforcement measures.

[0139] State VII (Critical State): Any location outside the layer. And model residuals This indicates a sudden change in the physical relationship between the anchor bolt force and the delamination displacement, which the model can no longer accurately describe. This is a precursor to overall instability, and work must be stopped and personnel evacuated immediately.

[0140] In this technical solution, the delamination location (excluding the free section / anchored section / anchor bolt), the comprehensive support performance index SEI, and the dynamic threshold are integrated. Based on the comparison results and the continuous changing trend of the optimal model parameters ("stable" or "continuously decreasing"), a state determination system with seven risk levels from low to high was constructed. This system not only covers different locations of the delamination from shallow to deep and the gradual deterioration process of the anchor bolt's working performance, but also specifically sets up a system based on model residual anomalies (…). The system can identify the critical state (state VII) of the physical relationship, which is a precursor to overall instability. Each state corresponds to a clear mechanical meaning and engineering warning signal, which effectively solves the problems of simple logic, easy false alarm and difficulty in distinguishing failure modes in traditional single-indicator alarms. It provides a step-by-step and operable decision-making basis for the site, from strengthening observation to immediate shutdown and evacuation.

[0141] Step 8: Output of tiered early warning and response recommendations;

[0142] Based on the determined status, corresponding warning levels and handling suggestions are output and pushed to relevant personnel and equipment in real time through various communication methods;

[0143] As a preferred approach, the process for outputting tiered early warning and response recommendations is as follows:

[0144] Based on the status determined in step 7, the following warning level and handling recommendations are output:

[0145] When in State I, a Level I warning (blue) will be issued, along with a reminder to maintain normal monitoring.

[0146] When the status is II, a Level II warning (yellow) will be issued, along with a reminder to strengthen monitoring and, if necessary, install additional anchor bolts.

[0147] When the status is III, a Level II warning (yellow) will be issued, and a reminder message will be sent to increase the intensity of monitoring and pay attention to changes in the anchorage section;

[0148] When the status is IV, a Level III warning (orange) will be issued, along with a reminder message to reinforce the anchorage section with grouting or to install additional long anchor cables.

[0149] When the status is V, a Level III warning (orange) will be issued, along with a reminder to strengthen deep support and consider installing additional anchor cables;

[0150] When the status is VI, a Level IV warning (red) will be issued, along with a reminder to immediately take deep reinforcement measures and prepare for production shutdown;

[0151] When status VII is reached, a Level V warning (red) will be issued, along with a reminder to immediately stop work, evacuate personnel, and fully reinforce the structure.

[0152] Meanwhile, the early warning information is pushed to the wellhead dispatch center and the mobile terminals of on-site personnel in real time, and the audible and visual alarms set up in key locations underground are used to carry out early warning actions.

[0153] This technical solution establishes a five-level progressive early warning system, from Level I (blue) to Level V (red), based on seven status determination results. Each level corresponds to clear and actionable engineering response suggestions (such as maintaining the status quo, strengthening observation, grouting reinforcement, work stoppage and evacuation, etc.), achieving precise matching between early warning levels and on-site response measures. Simultaneously, early warning information is pushed out in parallel through three channels: the surface dispatch center, on-site personnel mobile terminals, and underground audible and visual alarms, ensuring the real-time nature, redundancy, and coverage of the early warning. This effectively solves the problems of delayed early warning information transmission and vague response suggestions in traditional monitoring, providing a complete execution chain from "reminder" to "mandatory evacuation" for mine roof disaster prevention and control.

[0154] Step 9: Online self-learning of model parameters;

[0155] The judgment results and corresponding data after manual review are stored in the historical database, and key parameters are refitted periodically to achieve dynamic adaptation of the evaluation method to changes in geological conditions.

[0156] In this way, by introducing an online self-learning mechanism for model parameters, the method acquires the ability to dynamically adapt to geological conditions.

[0157] As a preferred option, the online self-learning process of model parameters is as follows:

[0158] S91: Data entry; After each manual review and confirmation of the accuracy of the status determination result, the data and inversion parameters in this window are stored in the historical database;

[0159] S92: Update: After every 30 sets of manually verified data, refit the surrounding rock rheological attenuation coefficient. The baseline value (obtained from the inversion of each window) (The median or average), which can be used as the benchmark value for subsequent inversion in a new window. The initial or fixed value;

[0160] S93: Update: The design reference force transmission coefficients for different surrounding rock categories are updated every 100 data sets. Value table (obtained from actual inversion) (Statistical corrections were made based on the corresponding surrounding rock conditions).

[0161] S94: Effect; Through the above-mentioned online self-learning mechanism, the evaluation method can adapt to the gradual or sudden changes in the properties of the surrounding rock during the tunnel excavation process.

[0162] In this technical solution, the judgment results after manual verification and confirmation are stored in a historical database along with the inversion parameters, and the rheological attenuation coefficient of the surrounding rock is periodically refitted using the accumulated data. (Every 30 groups) and updated design reference force transmission coefficients for different surrounding rock categories (Per 100 groups) This mechanism enables the dynamic adaptive adjustment of key model parameters to gradual or abrupt changes in geological conditions. This mechanism allows the evaluation method to continuously learn from field data, effectively avoiding the problem of fixed parameter models gradually becoming inaccurate due to changes in surrounding rock properties during long-term tunnel excavation. It significantly improves the time adaptability and geological universality of the method.

[0163] This invention covers the stages of cross-section layout and data acquisition, preprocessing and feature calculation, multi-model inversion and delamination location, SEI construction and dynamic threshold calibration, state determination and graded early warning, and online parameter self-learning. Each stage is clearly connected and can be implemented using only existing mine delamination instruments, anchor bolt force gauges, and ring network communication. No new dedicated hardware is required. The entire set of early warning information is pushed through multiple levels via the dispatch center, mobile terminals, and underground audible and visual alarms, facilitating rapid on-site response.

[0164] Comparative experiments to verify:

[0165] To verify the effectiveness of the method of this invention, a typical monitoring section was selected in a mine's main transport roadway for 60 days of continuous monitoring. The surrounding rock of this section was siltstone with well-developed fractures, and the anchor bolt specifications were Φ22×2400mm with a spacing of 800×800mm. During the monitoring period, three stages were successively experienced: the normal stage (days 1-30), the tunneling disturbance stage (days 31-45), and the instability precursor stage (days 46-50, with a collapse occurring on day 50).

[0166] Simultaneously deploy the four schemes listed in Table 1 below for comparison:

[0167] Table 1: Comparison of Schemes

[0168]

[0169] I. Normal Phase (Days 1-30);

[0170] The delamination amount fluctuated between 8 and 22 mm, and the anchor bolt axial force fluctuated between 45 and 72 kN, showing a consistent trend. The multi-model inversion results of this invention show that Model A has the smallest residual ( , , ), and determined that the separation layer is located in the free segment; inversion yielded N / m, Dynamic threshold > 0.81 The results were continuous and stable, classifying it as State I (low risk). None of the four schemes produced false alarms, but this invention additionally outputs the delamination location and quantitative performance evaluation.

[0171] II. Disturbance Deterioration Phase (Days 31-45);

[0172] On day 35, a tunneling disturbance occurred, with the delamination amount rapidly increasing from 18 mm to 41 mm and the axial force increasing from 58 kN to 98 kN before stabilizing. The tracking window of this invention was automatically reset and switched to a 5-minute short window. The multi-model inversion results for the stabilization period after the disturbance (day 45) are shown in Table 2.

[0173] Table 2 Comparison of Multi-Model Inversion Results (Table 1)

[0174]

[0175] The delamination has extended from the free section to the anchored section. The inversion yielded... N / m (design value) ), <Dynamic threshold 0.84, and If the price decrease is greater than 10% for three consecutive windows, the system is classified as State IV (higher risk), and a Level III orange alert is issued. A comparison of the results for each scheme is shown in Table 3.

[0176] Table 3 Comparison of Results for Each Scheme

[0177]

[0178] On-site drilling confirmed the presence of obvious delamination cracks at the interface of the anchorage section, verifying the correctness of the invention's determination.

[0179] III. The pre-instability stage (days 46-50);

[0180] An anomaly occurred on day 48: the delamination amount increased rapidly from 35 mm to 78 mm within 2 hours, but the anchor bolt axial force only increased slightly from 82 kN to 89 kN and then stopped increasing, indicating a "disconnection" between the delamination and the axial force. The inversion results of the three models of this invention are shown in Table 4:

[0181] Table 4 Comparison of Multi-Model Inversion Results (Table 2)

[0182]

[0183] The residuals of all three models were significantly larger than expected, and All values ​​close to or exceeding 1 indicate that the model fitting residuals are on the same order of magnitude as the fluctuation of the anchor bolt axial force itself, and the physical relationship between the anchor bolt force and the delamination displacement has undergone a fundamental abrupt change. The minimum residual is model C, indicating that the delamination has extended beyond the anchor bolt section. The calculation shows that 0.026 → 0, and SEI approaches zero, while simultaneously satisfying... The failure criterion for →1 triggers state VII (critical state), outputting a level V red alert, recommending immediate work stoppage and evacuation. A comparison of the results for each scheme is shown in Table 5.

[0184] Table 5 Comparison of Results for Each Scheme (Part 2)

[0185]

[0186] The collapse occurred approximately 20 minutes after the warning was issued, and the 15-minute lead time was sufficient for personnel evacuation. Comparative examples 1 and 3, which relied on the over-limit alarm for separation, showed a significant lag, while comparative example 2 completely missed the alarm because the axial force no longer increased.

[0187] IV. Overall Conclusion;

[0188] The results of 60 days of continuous monitoring show that the method of this invention:

[0189] 1. Disturbance degradation detection rate: 100%, compared to only 33% for the control group;

[0190] 2. Early warning lead time for instability precursors: approximately 15 minutes earlier than the threshold method;

[0191] 3. Delamination positioning capability: It can identify sections other than free sections / anchored sections / anchor bolts, which is not available in the comparison model;

[0192] 4. False alarm rate: 0%, consistent with the comparative rate;

[0193] 5. Output information richness: Provides quantitative SEI indicators, specific delamination locations, and graded treatment recommendations, which is significantly better than the comparative example.

[0194] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for intelligent evaluation of the anchorage status of roadway roof, characterized in that, Includes the following steps: Step 1: Joint monitoring section layout and data acquisition; synchronously acquire data on top plate delamination displacement and anchor bolt axial force within the same monitoring section; Step 2: Data preprocessing and feature calculation; The data is preprocessed to calculate the delamination and axial force increment rates using an adaptive window, and the window is automatically reset during strong disturbances. Step 3: Constructing multiple anchor bolt force models based on differences in mechanical response; Divide the top plate along the depth into a free section, an anchored section, and a section outside the anchor bolt, and establish anchor bolt force models for the three assumed separation locations respectively; S31: Based on the structural characteristics of the anchor bolt, the surrounding rock of the roof is divided into three sections along the depth direction: the free section, the anchored section, and the section outside the anchor bolt. S32: For the three sections where delamination may occur, establish anchor bolt force models under the assumed delamination locations: Model A is constructed based on the following formula: ; Model B is constructed based on the following formula: ; Model C is constructed based on the following formula: ; In the formula, , where is the delamination rate; The elastic modulus of the anchor bolt; This represents the cross-sectional area of ​​the anchor bolt. This refers to the length of the free section of the anchor bolt. This is the vertical distance from the delamination zone to the anchoring end of the anchor bolt; This is the total length of the anchor bolt; The equivalent interaction coefficient of the free segment; Shear stiffness at the anchorage section interface; This refers to the equivalent force transmission stiffness between the end of the anchor bolt and the deep delamination zone. The rheological attenuation coefficient of the surrounding rock; Step 4: Delamination location determination based on goodness-of-fit comparison; within the same sliding window, use continuous measured data to fit three different anchor bolt force models, determine the delamination location, and output the optimal parameters; Step 5: Construction of the comprehensive support effectiveness index; Based on the optimal model parameters and fitting residuals, construct the comprehensive support effectiveness index SEI; For cases where the delamination is located within the free section or anchorage section, the comprehensive index of support effectiveness should be constructed according to the following formula. : ； In the formula, Pick or ; Fit the residual standard deviation to the corresponding model; The standard deviation of the measured values ​​of the anchor bolt axial force within the window; Used as a reference for shear stiffness in design; For sections where the delamination is located outside the anchor bolts, the comprehensive index of support effectiveness should be constructed according to the following formula. : ； In the formula, For design reference force transmission coefficient; Step 6: Dynamic threshold calibration; Calculate dynamic thresholds based on SEI sequences from historical normal support phases; Step 7: Determine the support status; By combining the location of the delamination, the comparison between the current SEI and the dynamic threshold, and the trend of parameter changes, the support status is divided into multiple risk levels; Step 8: Output of tiered early warning and response recommendations; Output the corresponding warning level and handling recommendations.

2. The intelligent evaluation method for the anchorage status of roadway roof as described in claim 1, characterized in that, In step 1, the process of joint monitoring section layout and data acquisition is as follows: S11: Within the same roadway roof monitoring section, the roof separation instrument is installed in the roof separation instrument borehole in the middle of the roof, and the anchor bolt force gauge is installed on the adjacent anchor bolt trays on both sides of the separation instrument borehole. The horizontal distance between the adjacent anchor bolts and the separation instrument borehole is less than 0.5m. S12: Data is collected simultaneously using a roof separation instrument and a bolt force gauge, and the data is first stored in the downhole monitoring substation; S13: Data from underground monitoring substations is transmitted to the ground server in real time via a mining ring network; S14: The top plate delamination meter outputs the relative delamination displacement of the shallow base point relative to the deep base point. The anchor bolt force gauge outputs the axial force of the anchor bolt. .

3. The intelligent evaluation method for the anchorage status of roadway roof as described in claim 1, characterized in that, In step 2, the data preprocessing and feature calculation process is as follows: S21: Regarding the relative delamination displacement and anchor bolt axial force The original sequence is subjected to low-pass filtering for noise reduction and linear interpolation for missing values; S22: Define the length of the sliding time window The frequency of tunneling disturbance is selected within the range of 5 to 15 minutes. Specifically, 5 minutes is used when the tunneling cycle advance is large or the surrounding rock deformation rate is high, and 15 minutes is used in stable sections. S23: Calculate the delamination rate within each window. and the rate of increase of anchor bolt axial force ; S24: When strong disturbances occur, such as when the tunneling machine is cutting hard rock, the sliding window is automatically reset to eliminate sudden interference.

4. The intelligent evaluation method for the anchorage status of roadway roof as described in claim 1, characterized in that, In step 4, the process of determining the delamination location based on the goodness-of-fit comparison is as follows: For each sliding window, take all consecutive values ​​within the window. The measured data at each moment ,in Substitute the data into models A, B, and C respectively, and use the nonlinear least squares method to invert the unknown parameter sets of each model. Model A Model B Model C The objective function is the sum of squared residuals, as shown in the following equation: ； In the formula, ; The model predicts the anchor bolt axial force; the minimum residuals of the three models are compared. , , The delamination location represented by the model with the smallest residual is taken as the discrimination result of the current window: if The minimum delamination occurs in the free segment; if The minimum occurs in the anchorage section; if The minimum occurs in the section outside the anchor bolt; at the same time, the parameter estimates under the optimal model are recorded. , or And calculate the standard deviation of the fitted residuals. ,in This represents the number of parameters to be estimated in the model.

5. The intelligent evaluation method for the anchorage status of roadway roof according to claim 4, characterized in that, In step 4, the nonlinear least squares method employs the Levenberg-Marquardt algorithm, with the iteration termination condition being that the parameter change is less than 10. -6 Or the residual change is less than 10 -8 .

6. The intelligent evaluation method for the anchorage status of roadway roof according to claim 1, characterized in that, In step 6, the dynamic threshold calibration process is as follows: Collect the current moment before A sliding window Historical values, calculate their mean and standard deviation The dynamic threshold is defined according to the following formula. : ； History window length Adaptive adjustments based on the tunnel service stage: [During excavation] During the recovery period ; After each strong disturbance, the statistical parameters are recalculated; when the current window... At that time, the support effectiveness was determined to be significantly lower than the historical normal level.

7. The intelligent evaluation method for the anchorage status of roadway roof as described in claim 1, characterized in that, In step 7, the support status determination process is as follows: Based on the delamination location determined in step 4, the current... With dynamic threshold Based on the magnitude relationship and the changing trend of the optimal model parameters, the following seven states are defined: State I: The delamination is located within the free segment. ,and Stablize; State II: Delamination is located within the free segment. ,and Continued decline; State III: Delamination is located within the anchorage section. ,and Stablize; State IV: Delamination is located within the anchorage section. ,and Continued decline; State V: Delamination is located outside the anchor bolt. ,and Stablize; State VI: Delamination is located outside the anchor bolt. ,and Continued decline; State VII: Any position outside the layer. And model residuals .

8. The intelligent evaluation method for the anchorage status of roadway roof according to claim 1, characterized in that, In step 8, the process for outputting tiered early warning and response recommendations is as follows: Based on the status determined in step 7, the following warning level and handling recommendations are output: When in State I, a Level I warning will be issued, along with a reminder to maintain normal monitoring. When the status is II, a Level II early warning will be issued, along with a reminder to strengthen monitoring and install additional anchor bolts; When the status is III, a Level II early warning will be issued, and a reminder message will be sent to increase the intensity of monitoring and pay attention to changes in the anchorage section; When the status is IV, a Level III warning will be issued, along with a reminder message to reinforce the anchorage section with grouting or to install additional long anchor cables; When the status is V, a Level III warning will be issued, along with a reminder to strengthen deep support and consider installing additional anchor cables; When the status is VI, a Level IV warning will be issued, along with a reminder to immediately take deep reinforcement measures and prepare for a production shutdown. When status VII is reached, a Level V warning will be issued, along with a reminder to immediately stop work, evacuate personnel, and fully reinforce the structure. Meanwhile, the early warning information is pushed to the wellhead dispatch center and the mobile terminals of on-site personnel in real time, and the audible and visual alarms set up in key locations underground are used to carry out early warning actions.

Images