A full life cycle intelligent perception anchor rod supporting system in a coal mine underground

Abstract

The present application relates to the technical field of coal mine roadway supporting, and discloses a coal mine underground full life cycle intelligent sensing anchor rod supporting system, which comprises an anchor rod body, a construction machine and a controller. The anchor rod body is integrated with a fiber sensing array and a sound wave transducer; the construction machine comprises an automatic drilling rig; the controller is in communication connection with the fiber sensing array, the sound wave transducer and the automatic drilling rig, is used for performing baseline calibration, extracting parameters to establish a state baseline, performing cross-hole detection, using drilling rig operation vibration to solve a cross-hole attenuation factor, performing active detection, decoupling received ultrasonic signals in combination with the cross-hole attenuation factor, calculating an interface debonding index, performing closed-loop control, fusing the cross-hole attenuation factor and the interface debonding index to evaluate the regional state, and correcting construction parameters and issuing them to the automatic drilling rig according to the same. The present application eliminates the interference of geological background attenuation on signals, and realizes accurate evaluation of anchor rod damage and dynamic closed-loop control of supporting operation.

Description

Technical Field

[0001] This invention relates to the field of coal mine roadway support technology, specifically to an intelligent sensing anchor bolt support system for the entire life cycle of underground coal mines. Background Technology

[0002] Coal mine underground roadways face complex surrounding rock geological environments, and rock bolt support is a fundamental means of maintaining the stability of the surrounding rock. During the service life of rock bolts, acquiring their working status and evaluating the support effect plays a crucial role in ensuring the safety of mining operations. Currently, the industry typically uses technologies such as ultrasonic or fiber optic sensing to monitor the condition of rock bolts, and judges the support status by analyzing the acquired signal attenuation or reflection characteristics.

[0003] However, in practical applications, due to the variable geological properties of the surrounding rock in the well, the detection signal is superimposed with the attenuation characteristics of the rock layer background when propagating in the medium. When processing the sensing signal, the existing monitoring system has difficulty distinguishing between the acoustic dissipation characteristics of the surrounding rock medium itself and the debonding damage generated at the anchor interface. This results in a state coupling problem when the system analyzes the signal, and the final calculated interface damage assessment result is easily affected by geological factors and produces errors.

[0004] Furthermore, existing anchor bolt support systems typically separate the construction and monitoring phases. During the initial drilling operation, the drilling mechanical parameters generated by automated drilling rigs and other equipment are not translated into geological assessment data. Additionally, when insufficient local support strength is detected during subsequent monitoring, there is a lack of a mechanism to feed the assessment results back to the front-end construction equipment. This data fragmentation prevents the system from establishing a correlation between construction parameters and anchor service status. Consequently, it cannot adjust drilling distribution and grouting pressure in real time based on changes in geological and support conditions, making dynamic feedback and closed-loop control of the entire support operation difficult. Summary of the Invention

[0005] The technical problem solved by this invention is that the geological properties of the surrounding rock in coal mines are complex and variable. Conventional anchor bolt support monitoring systems have difficulty distinguishing between the acoustic attenuation characteristics of the surrounding rock medium itself and the debonding damage of the anchor bolt interface, resulting in errors in the assessment of the support status. At the same time, existing systems have difficulty establishing an effective correlation between the drilling machine data during the construction phase and the acoustic and optical sensing data during the service phase, and cannot perform real-time dynamic feedback and closed-loop control of support parameters based on the coupled changes in geology and anchoring status.

[0006] To address the above problems, the present invention provides the following technical solution:

[0007] This invention provides a fully lifecycle intelligent sensing anchor bolt support system for underground coal mines, comprising an anchor bolt body, construction machinery, and an edge computing controller. The anchor bolt body integrates a fiber Bragg grating sensor array and a piezoelectric ceramic transducer. The construction machinery includes an automated drilling rig and an intelligent grouting pump. The edge computing controller is communicatively connected to the fiber Bragg grating sensor array, the piezoelectric ceramic transducer, and the automated drilling rig.

[0008] The edge computing controller is used to perform baseline calibration, cross-hole detection, active detection, and closed-loop control steps: In the baseline calibration step, mechanical parameters from the construction period and the initial acoustic-optical signals of the anchor bolt body are extracted to establish a state baseline. The state baseline includes the drilling dynamics baseline and the initial acoustic-optical state baseline. The specific process for establishing the drilling dynamics baseline includes: By integrating the instantaneous advance depth recorded by the displacement sensor, the drilling rig push pressure, slewing angular velocity, slewing torque and instantaneous advance speed obtained by the drilling-while-drilling sensing module are resampled and mapped to the same depth coordinate system for data alignment. Invalid records that are filtered out when the instantaneous advance speed is lower than the preset dead zone threshold or when the advance pressure drops sharply are removed. A cubic spline interpolation algorithm is used to process missing data segments to ensure the continuity of the spatial sequence. Substituting valid data at the same depth coordinates into the drilling specific energy calculation formula, the unit rock-breaking energy consumption value at each depth is calculated and stored in spatial sequence form as the drilling dynamics baseline. The specific process of establishing the initial acoustic-optical state baseline includes: The actual reflection center wavelength of the fiber Bragg grating sensor array after the initial preload is applied and temperature compensation is completed is obtained, and the initial static strain field is calculated by combining the factory calibration wavelength data. A high-frequency pulse excitation signal is sent to the piezoelectric ceramic transducer through the ring power supply contact to generate ultrasonic waves. The high-speed dynamic demodulator simultaneously captures the dynamic wavelength drift signal and performs square integration within a set time window to obtain the energy spatial distribution parameters of the initial transmitted wave. The initial static strain field is bound to the energy spatial distribution parameters of the initial transmitted wave, and encapsulated as an initial acousto-optic state baseline. This step enables the quantitative recording of the original stratum strength and initial stress state at the anchorage location.

[0009] In the cross-hole detection step, the mechanical vibrations generated by the automated drilling rig operating in adjacent holes are used as the detection sound source. Combined with the drilling dynamics baseline, the cross-hole attenuation factor is calculated by receiving the cross-hole acoustic emission signals. The processing and feature extraction of the cross-hole acoustic emission signals include: The high-frequency fluctuation component of the instantaneous rotational torque of the automated drilling rig during rock breaking operations is extracted as a discrete mechanical prior reference template sequence; Envelope detection is performed on the optical wavelength drift sequence captured by the fiber Bragg grating sensor array, and downsampling interpolation is performed based on the mechanical basic sampling frequency to obtain the time-aligned optical wavelength drift sequence at the receiver. Within a delay time window defined by the spatial rectilinear propagation distance, a sliding cross-correlation operation is performed between the discrete mechanical prior reference template sequence and the optical wavelength drift sequence at the receiving end; Peak search is performed on the cross-correlation function value sequence of the source and receiver signals, and the squared amplitude of the maximum peak value is extracted as the effective transmitted wave energy parameter after enhancement. The process of calculating the transaperture attenuation factor includes: The enhanced effective transmitted wave energy parameter is multiplied by the square of the linear propagation distance between the rock-breaking source and the grating node to calculate the equivalent wavefront energy after geometric diffusion compensation. The relative transmission energy ratio is calculated by dividing the equivalent wavefront energy by the initial mechanical radiation energy reference value and multiplying it by the unified calibration constant across the entire link. Based on the exponential decay model, a logarithmic transformation is performed on the relative transmission energy ratio to calculate the cross-hole attenuation factor connecting the corresponding spatial paths. This factor reflects the acoustic dissipation characteristics of the surrounding rock medium itself.

[0010] In the active detection step, the piezoelectric ceramic transducer is triggered to emit ultrasonic waves, which are received by a fiber Bragg grating sensor array. Using the initial acousto-optic state baseline as a reference, the interface decoupling index is calculated by combining the transpore attenuation factor to achieve state decoupling. The state decoupling process includes: The edge computing controller extracts the reflected wave sequence corresponding to the ultrasonic wave, and after DC bias removal and bandpass filtering, calculates the sum of squares of the amplitudes of the sampling points and divides it by the preset reference excitation energy constant to solve for the relative reflection energy ratio. Extract the reference value of the equivalent trans-aperture attenuation factor corresponding to the current monitoring node after spatial interpolation preprocessing; The measured relative reflection energy ratio and the equivalent trans-hole attenuation factor reference value were substituted into the nonlinear decoupling model to calculate the interface debinding index. The interface debonding index is calculated as the quotient of the relative reflection energy ratio divided by the acoustic impedance compensation term, which includes a negative exponential term of the equivalent cross-hole attenuation factor reference value. By introducing the cross-hole attenuation factor, the interference of geological background attenuation on the ultrasonic reflection signal is eliminated, enabling accurate assessment of the degree of debonding damage at the anchor bolt interface.

[0011] In the closed-loop control step, the cross-hole attenuation factor and the interface debonding index are integrated to assess the regional state. Based on this, the construction parameters are adjusted and sent to the automated drilling rig for closed-loop control. The process of assessing the regional state includes: The target support area is discretized in three-dimensional space using a regular voxel mesh; Using a three-dimensional inverse distance weighted interpolation algorithm, the trans-hole attenuation factor and the interface deadhesion index in the local sensor coordinate system are aligned to the corresponding three-dimensional voxel units; Extreme value normalization was performed on the transpore attenuation factor and the interface debonding index within the same three-dimensional voxel unit, and then linearly added together with the set fusion weight coefficient to calculate the regional comprehensive state index characterizing the health of local support. The regional comprehensive state indices of all three-dimensional voxel units are tensor-stitched according to spatial topological relationships to construct a regional state evaluation matrix. The process of correcting construction parameters and executing closed-loop control includes: Voxel units whose regional comprehensive state index exceeds the preset safety critical threshold are selected from the regional state assessment matrix and connected and delineated as weak geological zones using spatial density clustering algorithm. The normalized over-limit ratio is calculated based on the difference between the arithmetic mean of the comprehensive state index of the voxel unit region within the weak geological zone and the preset safety critical threshold. The normalized over-limit ratio is multiplied by the preset grouting pressure compensation coefficient and a constant 1 is added. Then, it is multiplied by the preset standard foundation grouting pressure constant to solve for the target dynamic grouting pressure. The standard industrial control message containing the three-dimensional coordinates of the reinforced borehole and the target dynamic grouting pressure is sent to the automated drilling rig and intelligent grouting pump to perform targeted densified drilling and dynamic pressure boosting grouting operations, thereby achieving closed-loop control of the support parameters.

[0012] Furthermore, the fiber Bragg grating sensor array is embedded in a groove on the surface of the anchor bolt body; the piezoelectric ceramic transducer is integrated into the end region of the anchor bolt body. The anchor bolt body is fixedly installed with a connector assembly at its tail. The connector assembly is provided with a beam-expanding magnetic photoelectric composite interface, and internally encapsulates a miniature collimating lens group and a ring power supply contact, and integrates a miniature drive control circuit. The miniature collimating lens group is connected to the fiber Bragg grating sensor array, and the annular power supply contact is electrically connected to the miniature drive control circuit and the piezoelectric ceramic transducer.

[0013] The power head of the automated drilling rig is equipped with a drilling-while-drilling sensing module, a displacement sensor, and a magnetic follow-up connector. The magnetic follow-up connector is connected to the beam-expanding magnetic photoelectric composite interface. The automated drilling rig's onboard platform is equipped with a high-speed dynamic demodulator; The high-speed dynamic demodulator acquires the optical signals of the fiber Bragg grating sensor array through the magnetic servo connector; The miniature collimating lens group is used to amplify the single-mode beam in parallel divergence, constructing an optical transmission channel that can tolerate some dust particle obstruction. This structure avoids damage to traditional fiber optic physical plug-in interfaces caused by underground dust and vibration, ensuring the stability of signal transmission under complex operating conditions.

[0014] This invention provides an intelligent sensing anchor bolt support system covering the entire lifecycle of coal mine operations. It offers the following advantages: 1. This invention compensates for ultrasonic reflection signals by introducing a cross-hole attenuation factor, decoupling the acoustic attenuation characteristics of the surrounding rock medium from the debonding damage of the anchor bolt interface. This invention uses the mechanical vibration generated by the automated drilling rig as the sound source to calculate the cross-hole attenuation factor, and substitutes this factor into the nonlinear decoupling model in the active detection step, eliminating the interference of geological background attenuation on the calculation of the interface debonding index, thereby improving the accuracy of anchor bolt interface damage assessment.

[0015] 2. This invention constructs a regional state assessment matrix by integrating the cross-hole attenuation factor and the interface debonding index, thereby realizing dynamic closed-loop control of support parameters. This invention uses a spatial density clustering algorithm to identify weak geological zones, calculates the target dynamic grouting pressure based on the excess ratio of the regional comprehensive state index, and sends control messages containing borehole coordinates and grouting pressure to automated drilling rigs and intelligent grouting pumps. This transforms conventional passive monitoring into targeted operational adjustments based on the coupled changes in geological and anchoring states, thereby improving the support quality under complex geological conditions.

[0016] 3. This invention ensures the stability of photoelectric signal transmission in the underground environment by setting up a beam-expanding magnetic photoelectric composite interface containing a micro collimating lens group. This invention uses a micro collimating lens group to amplify the single-mode beam in parallel divergence, constructing a non-contact optical dustproof transmission channel. This avoids signal interruption and device damage caused by dust adhesion and mechanical vibration in coal mines, which is a problem with traditional physical plug-in interfaces. This ensures the continuous acquisition of sensor data throughout the entire service life of the anchor bolt. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall system architecture of the present invention; Figure 2 This is a flowchart illustrating the overall workflow of the present invention; Figure 3 This is a comparative analysis diagram of the evolution of the interface debonding index during the service life of the present invention. Detailed Implementation

[0018] The technical solutions in 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.

[0019] See attached document Figure 1 , Figure 1 This is a schematic diagram of the overall system architecture according to an embodiment of the present invention. The present invention provides an intelligent sensing anchor bolt support system for the entire life cycle of coal mines, which may include: an anchor bolt body, a connection end assembly, construction machinery end equipment, and a communication network architecture.

[0020] A sensor node array is coaxially mounted on the anchor bolt body. The sensor node array comprises a fiber Bragg grating sensor array and a piezoelectric ceramic transducer. The fiber Bragg grating sensor array is embedded in a groove on the surface of the anchor bolt body. The piezoelectric ceramic transducer is integrated into the end region of the anchor bolt body.

[0021] The connector assembly is fixedly installed at the tail of the anchor bolt body. The connector assembly is equipped with a beam-expanding magnetic-optical composite interface. This interface internally encapsulates a miniature collimating lens group and a ring-shaped power supply contact, and the connector assembly also integrates a miniature drive control circuit. The miniature collimating lens group connects to the optical path of the fiber Bragg grating sensor array to amplify the beam diameter. The ring-shaped power supply contact is electrically connected to the miniature drive control circuit and the piezoelectric ceramic transducer.

[0022] The construction machinery includes an automated drilling rig. The power head of the automated drilling rig is equipped with a drilling-while-feeding (DWF) sensor module and a magnetically attached follower connector. The DWF sensor module is used to collect thrust, torque, and rotational speed parameters in real time during the drilling process. The magnetically attached follower connector interfaces with a beam-expanding magnetic-optical composite interface. The onboard platform of the automated drilling rig carries a high-speed dynamic demodulator and an edge computing controller. The high-speed dynamic demodulator acquires the optical signals from the fiber Bragg grating sensor array through the magnetically attached follower connector. The edge computing controller establishes a data connection with external nodes through an industrial Ethernet gateway in the communication network architecture.

[0023] See attached document Figure 2 , Figure 2 This is a general workflow diagram according to an embodiment of the present invention. The overall workflow of the intelligent sensing anchor bolt support system for the entire life cycle of coal mines according to the present invention may include: S100, Drilling Connection and Baseline Synchronous Calibration. The automated drilling rig propels the anchor bolt body into the borehole. The magnetic follower connector and the beam-expanding magnetic photoelectric composite interface are automatically aligned and connected. The edge computing controller extracts the mechanical parameters output by the drilling-while-drilling sensor module to establish the drilling dynamics baseline for the current hole position. After applying the initial preload, the piezoelectric ceramic transducer emits the initial guided wave, and the fiber Bragg grating sensor array collects the transmitted energy data to complete the initial baseline calibration of the acousto-optic state.

[0024] S200, Passive Detection and Damage Assessment in the Neighborhood of Holes. The automated drilling rig moves to an adjacent hole to perform drilling operations. The edge computing controller uses the mechanical vibrations generated by drilling and rock breaking as a sound source, activating the fiber Bragg grating sensor array in the installed anchor bolt body through the communication network architecture. The fiber Bragg grating sensor array passively receives the cross-hole acoustic emission signals penetrating the surrounding rock. The edge computing controller performs signal reconstruction and attenuation calculation to obtain the surrounding rock damage characteristics between nodes.

[0025] S300, active collaborative detection and state decoupling during service life. During the service life of the anchor bolt, the system triggers the piezoelectric ceramic transducer to emit high-frequency ultrasonic guided waves at a set cycle. The fiber Bragg grating sensor array synchronously captures the residual energy distribution of the guided waves and measures the static strain change. The edge computing controller calculates the interface debonding parameters based on the energy and strain verification logic, and determines the stress release of the surrounding rock and the interface anchoring failure state.

[0026] S400, regional state fusion and construction parameter closed-loop control. The edge computing controller aggregates the interface debonding parameters and surrounding rock damage characteristics of multiple anchor bodies within a specified spatial area to generate a multi-dimensional state assessment matrix. The edge computing controller calculates compensation parameters based on this multi-dimensional state assessment matrix and sends the corrected target preload torque and support mesh spacing commands to the automated drilling rig.

[0027] Regarding the aforementioned drilling connection and baseline synchronization calibration step S100, the drilling connection and baseline synchronization calibration process provided by this invention establishes a data channel and acquires underlying parameters through the physical docking of the connection end component and the construction machinery end equipment. In this embodiment, the process specifically includes the following steps: S110. Automatic docking of mechanical and optical paths. The power head of the automated drilling rig drives the anchor bolt body into the borehole. During the propulsion action, the magnetic follower connector at the front of the automated drilling rig makes centering contact with the beam-expanding magnetic photoelectric composite interface at the rear of the anchor bolt body. As a preferred method, the miniature collimating lens group inside the beam-expanding magnetic photoelectric composite interface parallelly diverges the single-mode beam emitted from the fiber Bragg grating sensor array, amplifying the beam diameter to a preset size (e.g., tens to hundreds of times the original fiber core diameter), thereby establishing an optical transmission channel that can tolerate a certain amount of dust particle obstruction. The receiving lens inside the magnetic follower connector captures the amplified beam and couples it to a high-speed dynamic demodulator. The ring power supply contact and the power supply terminal of the magnetic follower connector form a closed electrical circuit. As a physical component for realizing non-contact dustproof transmission of optical signals and explosion-proof connection of electrical signals, the beam-expanding magnetic photoelectric composite interface may include a metal shell with a permanent magnet on the outside, an internal fiber ceramic ferrule, and a plano-convex collimating lens element. For the polarity setting and end-face sealing structure of magnetic components, those skilled in the art can select and assemble them according to the downhole explosion-proof level. The mechanical packaging process is a well-known technology in this field and will not be described in detail here.

[0028] S120. Extract drilling dynamic parameters and establish a strength baseline. During the drilling process, the edge computing controller collects mechanical operating data from the drilling-while-drilling sensor module in real time. Generally, the mechanical energy consumed by the drilling rig when breaking rock is highly correlated with the mechanical strength of the surrounding rock itself. Based on this principle, the edge computing controller calculates the drilling specific energy according to the collected data to quantify the mechanical strength of the original rock in the current hole depth direction. The drilling specific energy is calculated using the following formula: ; In the formula, Indicates the depth of the borehole Drilling specific energy at the location; Indicates the depth of the automated drilling rig The propulsion pressure applied at the drilling rig can be characterized by obtaining the values ​​of the hydraulic sensors in the drilling rig's propulsion cylinders. This indicates the cross-sectional area of ​​the drill bit used in automated drilling rigs; Indicates the depth of the drill bit angular velocity at the point; This represents the instantaneous rotational torque recorded by the drilling sensing module. This parameter can be calculated from the pressure difference between the inlet and outlet oil of the spindle hydraulic motor or measured directly by the torque sensor. This represents the instantaneous drilling speed at that depth. The edge computing controller calculates the drilling energy ratio at each depth. The data are fitted into a spatial sequence and used as the initial mechanical reference data of the surrounding rock at the service location of the anchor bolt body.

[0029] S130. Perform acoustic-optical synchronous triggering and initial state calibration. The automated drilling rig completes the implantation of the anchor bolt body and applies the set initial preload. The power head remains in the docked state, and the edge computing controller sends a high-frequency pulse excitation signal to the piezoelectric ceramic transducer through the ring power supply contact. The piezoelectric ceramic transducer is excited and generates an ultrasonic signal that propagates along the axial direction of the anchor bolt body. From a physical perspective, ultrasonic waves, as a type of mechanical elastic wave, cause changes in the dynamic stress field of the local medium when propagating inside the bolt. This tiny stress wave action is applied to the fiber Bragg grating sensor array, causing an instantaneous change in the grating's elastic-optical effect and its physical period, which is then manifested in the optical dimension as a corresponding shift in the reflection center wavelength. The high-speed dynamic demodulator synchronously captures this wavelength shift signal through a magnetically attached follower connector and extracts the energy spatial distribution parameters of the initial transmitted wave. The edge computing controller records the initial static strain field output by the fiber Bragg grating sensor array when it is not subjected to ultrasonic excitation. Initial static strain field Energy spatial distribution parameters of the initial transmitted wave Together, they form the initial acousto-optic baseline of the anchor bolt body. Regarding the inverse piezoelectric effect mechanism of ultrasonic waves excited by piezoelectric ceramics and the optical principle of grating center wavelength demodulation, those skilled in the art can perform conventional implementations based on existing acoustic and photoelectric detection theories. The underlying physical mechanisms are well-known technologies in this field and will not be elaborated upon here.

[0030] After the baseline calibration is completed, the power head of the automated drilling rig retracts. The magnetic servo connector and the beam-expanding magnetic photoelectric composite interface are physically separated. After separation, the construction personnel or the servo robotic arm physically connects the beam-expanding magnetic photoelectric composite interface remaining at the end of the borehole to the downhole photoelectric transmission network and long-term power supply harness fixedly laid in the roadway. Subsequently, the beam-expanding magnetic photoelectric composite interface enters a long-term service standby state to ensure the energy supply and data channel for passive sensing and active detection during its service life.

[0031] Regarding step S120, which involves extracting drilling dynamic parameters and constructing a strength baseline, the edge computing controller uses dynamic data generated by the automated drilling rig during construction to construct a sequence of background parameters reflecting the geological characteristics of the current borehole location. The extraction and calculation process of the drilling dynamic baseline specifically includes sub-steps: S121. Spatiotemporal Conversion and Alignment of Multidimensional Drilling Parameters. During the drilling operation of the anchor bolt holes, the drilling-while-drilling sensor module continuously monitors the operating status of the power head. The data output by conventional sensors has a built-in timestamp attribute, while the characteristics of rock formations are inherently distributed with spatial depth. To address this difference in physical properties, the edge computing controller converts the mechanical parameters in the time domain into a parameter sequence in the spatial depth domain. By integrating the instantaneous drilling depth recorded by the displacement sensor, the system maps the drilling rig's feed pressure, slewing angular velocity, slewing torque, and instantaneous drilling speed to the same depth coordinate system.

[0032] As a preferred approach, the edge computing controller presets a basic depth step size (e.g., setting a sampling interval of 5 millimeters) based on the geological exploration accuracy requirements and the actual sampling rate of the sensors. Based on this depth step size, the system uses an equidistant interpolation algorithm to resample the original discrete data, ensuring that sensor data from different dimensions are aligned at the same spatial resolution, thus avoiding data misalignment due to sampling rate differences. If sensor communication packet loss occurs within a certain step size interval, resulting in data loss, the system uses adjacent valid historical data for linear estimation to maintain the integrity of the basic time series.

[0033] S122. Judgment and Filtering of Abnormal Mechanical Parameters. Due to the complex geological conditions and mechanical operations downhole, raw drilling data is often accompanied by pulse noise or invalid records generated by mechanical idling. To ensure the accuracy of subsequent energy calculations, the edge computing controller introduces a judgment mechanism with set thresholds to clean the aligned spatial data. The system monitors the instantaneous drilling speed and propulsion pressure values ​​in real time. When the instantaneous drilling speed is lower than the preset dead zone threshold, or when the propulsion pressure experiences a precipitous drop that does not conform to the characteristics of continuous rock breaking, the system determines that the drill bit is in a state of dry drilling or stuck.

[0034] In this embodiment, the preset dead zone threshold is obtained through an unloaded operation calibration program executed by the automated drilling rig before construction, and its value range is typically set between 5% and 10% of the normal average rock breaking speed; while the criterion for a cliff-like drop can be set as a pressure drop gradient exceeding three standard deviations of the normal operating pressure. The edge computing controller then marks the data in this depth range as interference data and removes it.

[0035] To address the spatial sequence data gaps left after removing invalid records, the edge computing controller employs a cubic spline interpolation algorithm to smooth the transition of valid values ​​on both sides of the breakpoint, ensuring the continuity of the data sequence throughout the entire hole depth. For conventional high-frequency noise removal, those skilled in the art can set up conventional algorithms such as median filtering based on the noise characteristics of the actual hardware. The basic data processing methods are well-known in the field and will not be elaborated upon here.

[0036] S123. Mapping and serialization of drilling dynamics baseline. After data cleaning and spatial alignment, the edge computing controller synchronously substitutes the effective propulsion pressure, slewing angular velocity, slewing torque, and instantaneous advance rate at the same depth coordinates into the drilling specific energy calculation formula described in the aforementioned embodiment. The system then sequentially calculates the unit rock-breaking energy consumption value for each discrete depth node on the borehole trajectory.

[0037] The calculated discrete data not only quantifies the initial mechanical strength of the surrounding rock at the borehole location from a macroscopic perspective, but also serves as a reference benchmark for the initial mechanical power of the seismic source during subsequent cross-hole transmission detection. To facilitate subsequent bidirectional collaborative decoupling and attenuation comparison, the edge computing controller combines the acquired drilling specific energy at each depth according to spatial coordinate order to construct a one-dimensional spatial sequence matrix. This sequence matrix is ​​persistently stored in local memory, formally completing the construction of the initial dynamic baseline of the surrounding rock at the service location of the anchor bolt.

[0038] Regarding step S130, which involves simultaneous acoustic and optical triggering and initial state calibration, after the automated drilling rig completes the implantation of the anchor bolt and applies the initial preload, the edge computing controller synchronously triggers the photoelectric sensing network to extract and solidify the acoustic and optical baseline data. This calibration process specifically includes the following sub-steps: S131. Non-destructive extraction of the initial static strain field. Typically, the anchor bolt body undergoes axial tensile deformation under preload. Considering the potential for geothermal activity deep within the mine and frictional heat generated during construction, which could cause thermal drift in the grating wavelength, as a preferred approach, the system introduces a temperature reference grating that operates in the same thermal environment as the strain grating but is not subjected to axial force for temperature compensation.

[0039] After the wavelength change induced by the stripping temperature, the edge computing controller obtains the actual reflection center wavelength of each node in the fiber Bragg grating sensor array after the initial preload is applied and temperature compensation is completed via a high-speed dynamic demodulator. Based on this parameter, the system combines the factory-calibrated wavelength data of each node under stress-free conditions to calculate the initial static strain field at the spatial location. This calculation process is achieved through the following formula: ; In the formula, Indicates spatial location First Initial static strain of each grating node; This indicates the actual reflection center wavelength of the grating node after the initial preload is applied and temperature compensation is completed; This represents the initial reflection center wavelength of the grating node in a free and stress-free state; This represents the effective elastic-optical coefficient of the fiber core. This parameter is determined by the matrix material properties of the fiber and is usually built into the system as a known constant. This static strain field can generally reflect the stress distribution in the initial stage of support and can serve as a zero-point for subsequent monitoring of abnormal stress release.

[0040] S132. Active Excitation of Ultrasonic Guided Waves and High-Frequency Optical Signal Capture. After the static baseline is recorded, the edge computing controller injects a high-frequency pulse voltage into the piezoelectric ceramic transducer through a closed loop. To obtain good electroacoustic conversion efficiency, the excitation frequency of this pulse voltage is usually set to match the mechanical resonant frequency of the piezoelectric ceramic transducer itself, and its value range is generally set between tens of kilohertz and several megahertz. The piezoelectric ceramic transducer generates ultrasonic elastic waves propagating along the anchor body when excited. When the stress wave passes through the grating node, it causes transient modulation of the fiber core refractive index and the physical period of the grating. The high-speed dynamic demodulator captures the dynamic wavelength drift signal sequence of each node at a preset high-frequency sampling rate. For the basic physical conversion relationship between the piezoelectric effect and the grating elastic-optical effect, those skilled in the art can deduce it based on existing physical sensing theories. The underlying transduction and modulation mechanism is a well-known technology in this field and will not be elaborated here.

[0041] S133. Integral Calculation and Baseline Matrix Encapsulation of Dynamic Energy Parameters. For the acquired high-frequency time-series signal, the edge computing controller performs energy dimension quantization. To ensure the accuracy of dynamic energy capture, the starting point of the integral calculation must be strictly synchronized at the hardware level with the electrical pulse triggering moment of the piezoelectric ceramic transducer. The system performs square integration on the dynamic wavelength drift amplitude within a set time window to obtain the initial transmitted wave's energy spatial distribution parameters. The specific calculation formula is as follows: ; In the formula, Indicates spatial location First The initial transmitted wave energy parameters obtained from each grating node; This indicates the set ultrasonic signal reception time window. The length of this time window is calculated and determined based on the propagation speed of ultrasonic waves in the metal material and the physical length of the anchor bolt body, to ensure that the reception period of the effective sound wave can be fully covered. This indicates that the grating node is in time. The captured instantaneous dynamic wavelength drift amplitude, where time It corresponds strictly to the excitation origin of the ultrasonic pulse.

[0042] After the calculation is completed, the edge computing controller will calculate the initial static strain of each node. With the initial transmitted wave energy parameter Data binding is performed. The system combines and encapsulates the data into an initial acoustic-optical state baseline matrix. This matrix serves as the basis for evaluating the acoustic-optical response characteristics of the anchor bolt body in the early stages of anchorage debonding without significant surrounding rock damage, and is stored in the database to provide a quantitative reference point for subsequent service life state evolution assessment.

[0043] In step S200, which involves passive detection and damage calculation across borehole neighborhoods, the edge computing controller utilizes the mechanical vibrations generated by the automated drilling rig during construction at adjacent borehole locations as a natural sound source for ground exploration. Combined with the photoelectric sensing network embedded in the anchor bolts, this enables collaborative detection of the surrounding rock conditions between boreholes. This triggering and receiving process specifically includes the following sub-steps: S210. Dynamic Position Tracking and Parameter Extraction of Drilling Vibration Source. When the automated drilling rig moves to an adjacent hole to be supported (defined as the excitation hole in this embodiment) for drilling, the drill bit cutting the rock strata will generate continuous mechanical rock-breaking vibrations. From the perspective of rock-breaking mechanism, the local compression and shearing action of the drill bit on the rock at the bottom of the hole will lead to the initiation and propagation of microcracks, and the instantaneous release of strain energy constitutes a broadband elastic stress wave radiating to the surrounding rock mass. During this operation, the edge computing controller obtains the instantaneous depth coordinates of the drill bit in the excitation hole in real time through the drilling-while-drilling sensing module. .

[0044] As a preferred approach, the system combines the drilling specific energy at that depth calculated in the aforementioned embodiments with a set acoustic emission conversion efficiency coefficient as the initial mechanical radiation energy reference value for the acoustic emission source. This acoustic emission conversion efficiency coefficient is typically associated with the brittleness index of the rock itself and can be set using prior data from regional geological exploration. This method of converting rock-breaking energy into detection energy input eliminates the cumbersome process of deploying additional artificial seismic sources downhole.

[0045] S220. Passive capture and photoelectric demodulation of cross-hole elastic waves. The radiated elastic stress wave penetrates the surrounding rock medium between the excitation hole and the supported hole (defined as the monitoring hole in this embodiment), reaching the anchor bolt body inside the monitoring hole, which has completed state calibration. The dynamic disturbance of the stress wave acts on the fiber Bragg grating sensor array inside the anchor bolt, causing transient modulation of the physical period of each grating node. At this time, the beam-expanding magnetic photoelectric composite interface at the tail of the monitoring hole maintains a closed communication loop with the high-speed dynamic demodulator through the downhole optical cable network or multi-channel relay equipment. The high-speed dynamic demodulator continuously records the reflection center wavelength drift signal caused by the broadband stress wave at a preset high-frequency sampling rate.

[0046] To effectively extract the characteristic wavebands containing valuable geological information from the massive amounts of sensing data generated by continuous mechanical rock-breaking operations, the edge computing controller introduces a Short-Time Average to Long-Time Average (STA / LTA) algorithm. When the energy ratio of the real-time wavelength drift sequence exceeds a preset trigger threshold, the system automatically extracts data within a set time period before and after the characteristic wave peak as a valid detection signal. For the propagation laws of broadband stress waves in soil and rock media and the fundamental theories of grating-sensing dynamic strain, those skilled in the art can refer to relevant rock mechanics and acoustics literature; the wave propagation and sensing characteristics are well-known technologies in this field and will not be elaborated upon here.

[0047] S230: Spatiotemporal synchronization and alignment of monitoring data with seismic source parameters. For acoustic emission excitation and photoelectric reception behaviors distributed in different spatial locations, the edge computing controller introduces a global clock synchronization mechanism to achieve spatiotemporal coordinate alignment between seismic source parameters and passively received signals. The system, based on a network timing protocol combined with underlying hardware trigger markers, accurately records the excitation aperture's depth... The moment of rock breaking action was recorded, and the data from the monitoring borehole was extracted. The effective optical wavelength drift sequence captured by each grating node within the corresponding delay time window.

[0048] To quantify the subsequent geometric diffusion attenuation, the edge computing controller further calculates the spatial rectilinear propagation distance between the seismic source location and each receiving node. Specifically, based on spatial geometric relationships, the system calculates the horizontal design aperture spacing between the excitation aperture and the monitoring aperture. The square of, and the first in the monitoring well Spatial depth coordinates of each grating node along the axis Same as the instantaneous depth coordinates of the current acoustic emission source in the excitation hole The depth in the excitation hole is calculated by taking the square root of the sum of the squares of the two differences. The rock-breaking source and the monitoring well at the location The straight-line propagation distance between each grating node .

[0049] As an optimized implementation method, if borehole inclination data is available at the construction site, the system can use the actually measured borehole deflection angle to adjust the horizontal design borehole spacing. Dynamic corrections are performed in a three-dimensional coordinate system to obtain a more reliable distance reference.

[0050] Calculated linear propagation distance This not only provides a geometric reference input for subsequent energy attenuation compensation, but also helps establish a reliable spatial mapping relationship between the received grating dynamic response signal and the drilling rig's working face. The edge computing controller stores the spatiotemporally aligned distance parameters and the corresponding wavelength drift sequence in a cache, ready to perform further attenuation factor calculations.

[0051] After completing the aforementioned spatiotemporal synchronization alignment, considering the background noise caused by the operation of electromechanical equipment and dynamic adjustment of ground stress in the deep mine, the received cross-hole acoustic emission signal is often masked by cluttered broadband interference. To extract effective microseismic features from the low signal-to-noise ratio environment, the edge computing controller introduces prior mechanical information from the source end to perform cross-correlation denoising and feature demodulation on the passive wavelength drift sequence captured by the grating. The algorithm process specifically includes the following steps: S240. Extraction and temporal resampling alignment of the mechanical prior reference template. During rock breaking operations performed by the automated drilling rig, the drilling-while-working sensor module synchronously records the mechanical load fluctuations of the power head. Considering the co-modulation characteristics of mechanical cutting vibration and outwardly radiated acoustic emission stress waves, the edge computing controller extracts the high-frequency fluctuation component of the instantaneous rotational torque output by the drilling-while-working sensor module, and normalizes it after filtering out the DC bias component caused by the constant propulsion of the drilling rig. This normalized component serves as the discrete mechanical prior reference template sequence. Given the physical reality that there is an order-of-magnitude difference between the mechanical sampling rate of the drilling sensor module and the optical sampling rate of the high-speed dynamic demodulator, directly performing sequence matching would lead to time scale misalignment.

[0052] To address this issue, as a preferred approach, the edge computing controller performs envelope detection on the optical wavelength drift sequence to extract its low-frequency modulation envelope, and then performs downsampling interpolation on this envelope signal based on the fundamental sampling frequency of the mechanical sensor to obtain the receiver's optical wavelength drift sequence after envelope extraction and resampling alignment. This allows for sequence alignment of source and receiver signals at the same time resolution scale. For the specific implementation of the envelope detection and resampling algorithm, those skilled in the art can refer to relevant digital signal processing manuals; its underlying mathematical derivation is well-known in the field and will not be elaborated upon here.

[0053] S250, cross-correlation demodulation calculation of source and receiver signals. After sequence alignment, the edge computing controller, within a set time delay window, performs discrete mechanical prior reference template sequence... Optical wavelength drift sequence at the receiver Perform sliding cross-correlation. Since signals from the same source exhibit a certain degree of correlation, while environmental random noise and mechanical rock-breaking actions are usually independent, cross-correlation can generally effectively amplify the components from the same source and suppress random noise. Specifically, the system calculates the discrete mechanical prior reference template sequence. Optical wavelength drift sequence at the receiver relative time delay The sum of sliding products is used to calculate the cross-correlation function value of the source and receiver signals. .

[0054] To avoid computational redundancy caused by global traversal calculations, the edge computing controller combines the spatial rectilinear propagation distance calculated in the preceding steps with the estimated elastic wave velocity of the rocks in the survey area to define a reasonable physical delay calculation range, thus reducing the relative time delay. The estimation is only iterated within this estimation window. The estimated elastic wave velocity can be obtained by referring to prior geological exploration reports of similar rock formations.

[0055] S260, Extraction of effective propagation features and transmission energy mapping. With different relative time delays... The source-receiver signal cross-correlation function value The sequence is generated, and the edge computing controller performs peak search on the sequence. To avoid misjudging random mathematical fluctuations caused by pure environmental noise as valid signals, the system sets a signal-to-noise ratio confidence threshold, which is typically set to 3 to 5 times the mean cross-correlation value during the background silence period. When the maximum peak value in the cross-correlation sequence exceeds this confidence threshold, the time delay corresponding to the maximum peak value is determined. It can be used to characterize the propagation time of acoustic emission stress waves excited by mechanical rock breaking through the surrounding rock medium between holes to reach the monitoring hole.

[0056] Simultaneously, the system extracts the squared amplitude of the maximum peak value as the enhanced effective transmitted wave energy parameter. This energy parameter largely eliminates interference from complex environmental noise, reflecting the residual energy level of the elastic wave after absorption and scattering by the rock medium. The edge computing controller packages and encapsulates the effective transmitted wave energy parameters extracted by each receiving node with spatiotemporal parameters, serving as a reliable data source input for subsequent construction of spatial attenuation profiles and assessment of surrounding rock damage.

[0057] After cross-correlation denoising and feature demodulation are completed and the spatiotemporally aligned transmitted wave energy parameters are obtained, the system further removes interference from physical spatial distance and source fluctuations to extract intrinsic attenuation characteristics that can objectively reflect the degree of damage to the surrounding rock medium between boreholes. The quantitative calculation process of this cross-bore attenuation factor (CAF) specifically includes the following steps: S270. Geometric Diffusion Compensation Based on Spatial Straight-Line Distance. When elastic stress waves propagate in rock media, their wavefront expands outward with increasing propagation distance, causing the wave energy per unit area to gradually decrease according to geometric laws. To eliminate this energy attenuation caused by spatial geometric expansion, the edge computing controller uses the spatial straight-line propagation distance calculated in the aforementioned steps to enhance the effective transmitted wave energy parameters extracted at the receiving end. Perform geometric diffusion compensation.

[0058] Specifically, the system will enhance the effective transmitted wave energy parameter Same as the rock-breaking epicenter The straight-line propagation distance between each grating node Multiply the squared terms to calculate the depth in the excitation hole. The rock-breaking seismic source was located at the monitoring borehole. Equivalent wavefront energy at each grating node after geometric diffusion compensation .

[0059] As a preferred approach, if the target area exhibits obvious layered rock mass characteristics, causing stress waves to diverge in a cylindrical rather than spherical manner, the system can adjust the exponential term of the distance from square to first power to adapt to the specific wave dynamics diffusion law.

[0060] S280. Dimensional Unification and Normalization of Source Mechanical Radiation Energy. During construction operations, the cutting resistance of automated drilling rigs changes dynamically as they traverse rock layers of varying hardness, causing fluctuations in the initial mechanical radiation energy emitted outwards. To ensure comparability of equivalent wavefront energies measured at different depths, the edge computing controller performs source normalization processing. Considering that the initial mechanical radiation energy of the acoustic emission source and the optical characteristic energy of the receiver belong to different physical dimensions, the system introduces a unified calibration constant across the entire link. Dimensional conversion is performed. Furthermore, considering that the initial mechanical radiation energy of the drilling rig may be extremely small when traversing cavities or extremely soft coal seams, directly substituting it into the denominator would cause calculation overflow.

[0061] Therefore, the system presets a minimum effective excitation energy threshold, which can be set based on the drilling rig's no-load power baseline. When the actual mechanical energy is lower than this threshold, the system directly discards the data for that depth node. Under the premise of meeting the energy threshold condition, the system will use the equivalent wavefront energy after geometric diffusion compensation. Divide by the initial mechanical radiation energy reference value of the acoustic emission source The resulting quotient is then multiplied by the end-to-end uniform scaling constant. This allows us to calculate the relative transmission energy ratio under the current spatial propagation path. .

[0062] Unified calibration constants across the entire link This calculation integrates the drill bit's transducer, the grating's photoelastic-optical conversion coefficient, and the instrument's amplification gain. It is typically determined through comparative measurements and experiments on known intact rock blocks during system factory calibration or the initial trial drilling phase of an engineering project. This step eliminates the interference of source-end excitation energy fluctuations on the detection results.

[0063] S290, logarithmic solution and physical mapping of the transaperture attenuation factor (CAF). After excluding geometric diffusion and source fluctuations, the attenuation of the relative transmitted energy ratio largely depends on the intrinsic absorption and microcrack scattering effects of the surrounding rock medium penetrated by the stress wave. The edge computing controller is based on the exponential attenuation model of waves in viscoelastic media, and calculates the relative transmitted energy ratio under the current spatial propagation path. Perform a logarithmic transformation to calculate a quantitative transpore attenuation factor. The factor is calculated using the following formula: ; In the formula, Indicates the depth of the connection excitation hole Location and monitoring hole The transaperture attenuation factor of the spatial path where each grating node is located; Indicates the rock-breaking source and the first The straight-line propagation distance between each grating node; This represents the relative transmission energy ratio under the current spatial propagation path; This represents a minimal positive real number compensation constant artificially set to prevent the logarithmic function from exhibiting mathematical singularities. Its value is usually set to one percent of the lower limit of the system noise-energy ratio.

[0064] In addition, considering that the heterogeneity of underground rock mass may lead to local acoustic focusing, causing the relative transmitted energy ratio to be briefly greater than 1 and thus calculate a negative attenuation factor, the system adds non-negative truncation logic here to force the possible negative values ​​to zero, so as to conform to the physical law of macroscopic energy dissipation.

[0065] The calculated cross-hole attenuation factor reflects the degree of discontinuity of the surrounding rock medium along the straight path in space. A smaller value indicates a relatively intact rock mass structure; a larger value indicates the possible presence of dense microcracks, joint zones, or loosened areas within the path region. The edge calculation controller aggregates the cross-hole attenuation factors from each intersecting path into a multi-dimensional array, establishing a rock mass damage attenuation characteristic matrix for the support section, thus providing a quantitative criterion for subsequent adaptive adjustment of support parameters.

[0066] During step S300 of active collaborative detection and state decoupling during service life, with the completion of anchor bolt installation and grouting, the support system enters its long-term service phase. At this point, the mechanical drilling rig has been removed. To continuously assess the hidden anchoring interface state, the edge computing controller utilizes an active acoustic network constructed from the anchor bolt internal components to implement collaborative detection of the interface state during service life. This process specifically includes the following sub-steps: S310, during service, the active ultrasonic anchor bolt issues detection commands. The built-in micro-drive control circuit responds to the commands, driving the piezoelectric ceramic transducer to emit high-frequency ultrasonic pulses of a specific frequency. To ensure that the detection beam can effectively identify interface debonding and micro-cracks, the center frequency of the pulse is usually selected according to the acoustic principle of matching the wavelength with the expected defect size, and its value is generally set in the ultrasonic frequency range of tens to hundreds of kilohertz.

[0067] To avoid weak excitation caused by aging of the piezoelectric ceramic or short circuit due to moisture in the drive circuit, the system incorporates low-level current monitoring logic in the micro-drive control circuit. When the measured drive current falls below the set normal operating limit, the system determines that the piezoelectric ceramic transducer at that node has physically failed and actively interrupts the subsequent calculation process for that node. This avoids misinterpreting a pure noise signal as a high transmission state, which could lead to a dead zone in the algorithm logic. Under normal excitation conditions, ultrasonic pulses radiate along the anchor bolt body outwards to the grout and surrounding rock medium, forming a known active seismic source for detection.

[0068] S320. Passive capture and timing interception of multiple reflection waves from interfaces. When radiated ultrasonic waves pass through the physical interfaces between the anchor body and the grouting body, and between the grouting body and the surrounding rock, the acoustic impedance mismatch caused by the differences in density and wave velocity between the different media usually produces a certain degree of waveform transmission and reflection. Fiber Bragg grating sensor arrays capture these reflected stress waves containing interface state information, causing a high-frequency transient drift in their center wavelength. A high-speed dynamic demodulator continuously records this optical response through a beam-expanding magnetic photoelectric composite interface. The edge computing controller, based on the active excitation time of the issued command, uses a fixed time window to intercept the reflected waveband from the continuous recording.

[0069] As a preferred approach, the start and end time parameters of this time window are set based on the theoretical round-trip travel time boundary of the ultrasonic wave in the grouting body of the designed thickness, with an additional appropriate time tolerance margin to accommodate minor fluctuations in wave velocity caused by uneven grouting layer thickness or geostress evolution. For the waveform reflection and transmission distribution in multilayer media, those skilled in the art can refer to basic acoustic theory; the interface reflection characteristics are well-known in the field and will not be elaborated upon here.

[0070] S330. Extraction of time-frequency domain feature parameters and mapping of reflected energy. For the intercepted reflected wave sequence, the edge computing controller performs DC bias removal processing to eliminate baseline drift interference caused by gradual changes in ground stress. Considering that broadband random interference may still exist in the mine service environment, directly calculating the sequence energy could easily lead to misinterpreting out-of-band noise as interface reflected energy. Therefore, the system introduces a bandpass filtering algorithm, ensuring that the passband range of the filter strictly corresponds to the center frequency emitted by the aforementioned piezoelectric ultrasonic excitation module, thereby effectively filtering out irrelevant frequency band noise.

[0071] After filtering and purification, the system extracts key parameters that characterize the physical state of the hidden interface. Specifically, the system calculates the discrete wavelength drift sequence within a set time window after truncation, DC removal, and bandpass filtering. The sum of the squares of the amplitudes at each sampling point is divided by the reference excitation energy constant corresponding to the piezoelectric ceramic transducer, which is entered into the system during the factory calibration phase. Thus, the number of monitoring holes can be calculated. The relative reflection energy ratio acquired by each grating node within the current detection period .

[0072] The calculated relative reflection energy ratio This largely reflects the bonding state between the grout and the surrounding rock. When the hidden interface is tightly bonded, a large amount of sound waves penetrate into the deep surrounding rock, and the received reflected wave energy remains at a low level. Once debonding, cracks, or voids caused by incomplete grouting occur in the interface area, the acoustic impedance mismatch at the interface intensifies, and the total reflection effect of ultrasound at the air interface often leads to a higher relative reflected energy. A significant numerical increase was observed. The edge computing controller sets the relative reflected energy ratio of each node. Saved to the database, providing direct physical state mapping input for subsequent bidirectional decoupling analysis and support failure early warning.

[0073] After extracting the time-frequency domain feature parameters, the edge computing controller, having obtained the relative reflection energy ratio during its service life, typically needs to handle multi-factor coupling interference. The energy level of the reflected wave is not only controlled by the actual debonding degree of the grout-to-surrounding rock interface, but also fluctuates due to changes in acoustic impedance caused by microcrack damage in the surrounding rock medium itself. To eliminate the masking effect of background damage in the surrounding rock on the reflection characteristics, the system establishes a two-way decoupling model of the interface debonding index (IDI), thereby achieving a quantitative assessment of the hidden interface state. This decoupling analysis process specifically includes the following steps: S340, spatiotemporal matching and interpolation preprocessing of multi-stage monitoring data. The edge computing controller calls the rock mass damage attenuation feature matrix generated in the database during the drilling stage. This is for the current phase of the operational detection... For each grating node, the system uses its three-dimensional physical coordinates as a reference and retrieves the cross-hole attenuation factor that matches the spatial location of the node from the rock mass damage attenuation feature matrix. Considering the heterogeneity of the geological body and the possible slight misalignment of node coordinates at different construction stages, the system uses a distance-inverse weighted algorithm to spatially interpolate the discrete attenuation data within the retrieval range, thereby obtaining the equivalent cross-hole attenuation factor reference value corresponding to the current node. .

[0074] To prevent the interpolation algorithm from collapsing in sparse data areas due to the inability to find neighboring nodes, the system presets a maximum effective search radius. When there is no effective attenuation data within this radius, the system automatically calls the historical average attenuation value of the same layer of rock at the depth of the monitoring node as a fallback compensation to ensure the continuity of the decoupled calculation process. For the underlying weight allocation theory of the spatial interpolation algorithm, those skilled in the art can refer to relevant basic theories in geostatistics, which are well-known techniques in this field and will not be elaborated upon here.

[0075] S350. Acoustic impedance compensation assessment based on background rock damage. The system uses the obtained equivalent trans-hole attenuation factor reference value corresponding to the current node. The macroscopic equivalent acoustic impedance attenuation effect of the surrounding rock was compensated for by calculation. From the perspective of acoustic detection principles, when there is a dense network of fractures within the surrounding rock, its macroscopic elastic modulus and medium density often show a decreasing trend, causing the actual acoustic impedance to deviate from the theoretical baseline of intact rock. This impedance mismatch means that even with good interface bonding, ultrasonic waves will still generate high background reflection energy at the physical interface between the relatively dense grout and the damaged surrounding rock due to the abrupt change in medium wave impedance.

[0076] The edge computing controller incorporates negative exponential smoothing mapping logic to calculate the reflectivity baseline offset caused by background damage to the surrounding rock. As a preferred approach, this baseline offset calculation logic extracts the equivalent cross-hole attenuation factor reference value corresponding to the current node. The negative exponent term is multiplied by a preset impedance reduction factor to achieve this, thereby avoiding the overcompensation problem caused by linear reduction in extremely fractured rock masses.

[0077] S360, Nonlinear Calculation and State Mapping of the Interface Debonding Index (IDI). After completing the acoustic impedance compensation assessment, the edge computing controller introduces a two-way decoupled physical model. The system nonlinearly fuses the measured relative reflection energy ratio with the surrounding rock acoustic impedance compensation term to remove background interference and calculate the interface debonding index, which purely reflects the physical separation state of the interface. This index is calculated using the following formula: ; In the formula, Indicates the monitoring well number The interface debonding index calculated at each grating node; Indicates the monitoring well number The relative reflection energy ratio of each grating node during the current detection period; This represents the reference value of the equivalent transaperture attenuation factor corresponding to the current node; This represents the interface impedance matching correction coefficient, whose value is strictly limited to the open interval between 0 and 1 to prevent physical anomalies and mathematical overflow dead zones where the denominator is less than or equal to zero. This represents the sensitivity constant for damage to the surrounding rock.

[0078] The two empirical constants mentioned above are usually calibrated and confirmed in the pilot test section of the target roadway through core sampling comparison experiments. Among them, the surrounding rock damage sensitivity constant... The value is mainly determined based on the attenuation gradient of the local rock strata elastic wave velocity with the fracture density.

[0079] Calculated interfacial deadhesion index It can objectively reflect the true bonding state between the grout and the surrounding rock interface. Values ​​approaching a specific baseline level indicate a relatively tight bond at the anchorage interface; abnormally large values ​​indicate potential substantial interface debonding or the presence of large-scale grouting voids in the area. The edge computing controller calculates the interface debonding index of all nodes. It is integrated into the support status assessment database to provide quantitative data support for subsequent determination of the overall stability of the roadway.

[0080] In step S400 of the regional state fusion and construction parameter closed-loop control, after the edge computing controller calculates the cross-hole attenuation factor reflecting the background damage of the surrounding rock and the interface debonding index reflecting the interface debonding state, it needs to comprehensively process the data distributed in different spatiotemporal dimensions to form a global state benchmark to guide the automated drilling rig in performing self-correction operations. The construction and spatiotemporal mapping process of this regional state assessment matrix specifically includes the following sub-steps: S410. Three-dimensional spatial meshing and data alignment based on the global absolute coordinate system of the roadway. Since the cross-hole attenuation factor and interface decoupling index often depend on different borehole exploration paths or service-life anchor array layouts when obtained, there are differences in their spatial distribution resolution and the location of the original data points. To achieve homogeneous comparison of multi-source data, the edge computing controller uniformly transforms the scattered data points in the local sensor coordinate system to the global absolute coordinate system of the roadway. The system uses a regular voxel mesh to spatially discretize the target support area, dividing the surrounding rock of the roadway into multiple three-dimensional voxel units of equal size.

[0081] As a preferred approach, the side length of the voxel unit is typically set based on the physical node spacing of the grating sensor array and the effective coverage radius of the ultrasonic detection, ensuring that meaningful physical information is reflected within each grid. For each voxel unit, the system uses a three-dimensional inverse distance weighted interpolation algorithm to interpolate and extract the transpore attenuation factor and interface decoupling index scattered in its neighborhood to the geometric center point of the voxel, thereby completing the alignment of heterogeneous parameters on a spatial scale.

[0082] During interpolation, if no valid data points are found within the preset initial search radius, the system will initiate dynamic search radius expansion logic. If no data is found even after reaching the maximum allowable search radius, the system will assign the prior safety mean of the geological stratum to the voxel to avoid computational dead zones caused by division by zero. For the construction of the underlying data structure of the 3D mesh and the solution of the spatial coordinate system transformation matrix, those skilled in the art can refer to relevant theories of computer graphics and spatial analytic geometry, which are well-known techniques in this field and will not be elaborated upon here.

[0083] S420, Dimensionless processing of multi-source parameters and calculation of regional comprehensive state index. After completing spatial grid alignment, since the surrounding rock damage characteristics and interface debonding characteristics belong to different physical dimensions and have large differences in numerical distribution, directly accumulating the values ​​will lead to distortion of the evaluation results.

[0084] The edge computing controller performs extreme value normalization on the two types of parameters contained in each voxel unit, mapping them to a dimensionless standard interval of zero to one. From a data processing perspective, this process involves subtracting the regional global minimum value from the current parameter value, then dividing by the difference between the regional global maximum and minimum values, thereby eliminating dimensional differences. Based on this, the system linearly fuses the normalized parameters by introducing information entropy weighting or engineering experience weighting coefficients to calculate the regional comprehensive state index characterizing the support health of the micro-region. .

[0085] Specifically, the calculation logic is to calculate the first value in the global coordinate system. A three-dimensional voxel unit at time node Equivalent trans-hole attenuation factor after normalization With the normalized interface deadhesion index Multiply by the fusion weighting coefficients assigned to the transpore attenuation factor and the interface debinding index, respectively. and Then, the values ​​are added together to achieve the desired result. The values ​​of the aforementioned fusion weighting coefficients all fall within a closed interval of 0 to 1, and the sum is usually limited to 1. In deep mines with a high tendency for ground stress impact, the weighting coefficient representing the background damage of the surrounding rock is usually appropriately increased. The value of is selected to enhance the early warning capability for rockburst risk.

[0086] S430, Dynamic Construction and 3D Visualization Mapping of the Regional State Assessment Matrix. The edge computing controller calculates the regional comprehensive state index from all 3D voxel units. Tensor splicing is performed based on the spatial topological relationships to construct a regional state assessment matrix describing the current support section. As the project progresses and the service life extends, the system will update the matrix at different time points. The generated regional state evaluation matrix is ​​stored according to temporal relationships, forming a four-dimensional spatiotemporal state array containing the time dimension.

[0087] This matrix data can be directly mapped onto a geological 3D model to generate a 3D state cloud map characterizing the distribution of support strength. In the cloud map mapping, areas with higher voxel values ​​typically correspond to weak zones with severely fractured rock mass and poor anchoring; areas with lower values ​​indicate a more stable bearing state of the support structure. This multidimensional parametric matrix not only provides maintenance personnel with an intuitive global state assessment view but also provides a standardized, machine-readable input source for subsequent automated drilling rigs to adaptively generate drilling and grouting correction strategies based on spatial weak points.

[0088] After constructing and mapping the aforementioned regional state assessment matrix, the edge computing controller, having completed the construction of the regional state assessment matrix containing spatiotemporal information, needs to convert the quantified assessment data into physical commands that can be directly invoked at the engineering level. The system guides the automated drilling rig to implement targeted reinforcement in subsequent working faces or to perform reinforcement grouting operations on the current support section through a dynamic adaptive adjustment strategy. The adaptive adjustment and execution process of these support parameters specifically includes the following steps: S440. Clustering and boundary delineation of local high-risk weak areas. The edge computing controller performs threshold traversal and filtering on the comprehensive state index of each three-dimensional voxel unit in the region state assessment matrix. When the comprehensive state index of a voxel exceeds the preset safety threshold, the system marks it as an abnormal unit. To avoid triggering a mathematical overflow dead zone with a denominator of zero in subsequent calculations, the safety threshold is strictly limited to an open interval of 0 to 1.

[0089] As a preferred method, calibration is usually performed by combining statistical data of historical support failure cases and pull-out test baselines of field test sections, with values ​​typically ranging from 0.6 to 0.8.

[0090] Furthermore, if no abnormal units exceeding the threshold are found during the traversal, the system will determine that the current support status is stable and directly use the basic design parameters to execute subsequent operations, thus avoiding the algorithm falling into a logical dead zone of infinite waiting due to the inability to find the target. Considering that mine rock mass damage and interface debonding often present as continuous sheet-like or band-like distributions, the system uses a spatial density clustering algorithm to perform spatial connectivity analysis on discrete abnormal units, merging physically close abnormal voxels into weak geological zones with clear geometric boundaries.

[0091] For the core parameter settings and underlying logic of connected component search in spatial density clustering algorithms, those skilled in the art can refer to the relevant basic theories of data mining, which are well-known technologies in this field and will not be elaborated here.

[0092] S450, Nonlinear Compensation and Dynamic Calculation of Key Construction Parameters. For the extracted weak geological zones, the system adaptively generates and adjusts the corresponding support parameters. This adjustment is mainly reflected in the increase of reinforcement borehole density and the targeted increase of grouting pressure. From the perspective of engineering fluid mechanics and the principles of soil and rock reinforcement, a high regional comprehensive state index often indicates the presence of dense microcracks or large-scale interfacial debonding voids. If conventional grouting pressure is maintained, the grout may solidify prematurely during infiltration due to excessive flow resistance, making it difficult to fully fill the microcracks. Therefore, it is necessary to moderately increase the grouting pressure to overcome the flow resistance of the grout and expand its effective diffusion radius within the loose zone.

[0093] As a preferred approach, the edge computing controller dynamically compensates for the foundation grouting pressure based on the average of the regional comprehensive state indices within the weak geological zone. Specifically, the calculation logic involves taking the arithmetic mean of the regional comprehensive state indices of all voxel units within the weak geological zone, extracted using a spatial density clustering algorithm. With the set safety threshold The difference between them, divided by a constant 1 and the set safety threshold, The difference between them is used to calculate the normalized out-of-limit ratio; Subsequently, the normalized excess ratio is multiplied by the grouting pressure compensation coefficient. Adding a constant of 1, the result is finally compared with the standard foundation grouting pressure constant designed for intact rock masses during the factory stage or early geological exploration. Multiplying these components allows for the calculation of the target dynamic grouting pressure output after adaptive adjustment. The above grouting pressure compensation coefficient It is mainly used to control the sensitivity of pressure increase, and its value is strictly limited to a closed interval of 0.1 to 0.5 to prevent excessive local grouting pressure from inducing secondary splitting damage of the surrounding rock.

[0094] S460, timing encapsulation of control commands and closed-loop execution at the device end. After completing the dynamic calculation of support parameters, the edge computing controller encapsulates information such as the target dynamic grouting pressure and the three-dimensional coordinates of the borehole to be encrypted into a standard industrial control message. This message is sent to the underlying programmable logic controller of the automated drilling rig via the downhole bus network. After parsing the command, the automated drilling rig drives the drilling rig attitude adjustment mechanism 36 set on the body to change the elevation and yaw angles of the drill arm, performing encrypted drilling operations in the actual physical space corresponding to the weak geological zone.

[0095] In the subsequent grouting process, the intelligent grouting pump 37, installed at the equipment end or in collaborative operation, outputs the target dynamic grouting pressure based on the received adaptive adjustment. The pump motor speed is adjusted in real time via a built-in frequency converter. This data-driven self-correcting control mechanism helps the grouting fluid penetrate into micro-cracks and debonding interfaces under pressure, thereby achieving a certain degree of compensation for hidden interface defects and closed-loop reinforcement of the support system.

[0096] Specific application examples: Application scenario setting: The 8304 return air roadway in a deep coal mine, buried at a depth of approximately 800 meters, exhibits high ground stress and contains localized concealed fault fracture zones. The roadway is supported using the intelligent sensing anchor bolt support system provided by this invention. The anchor bolt length is set at 2.5 meters, and the design foundation grouting pressure constant of the grouting body is... 2.0MPa.

[0097] Implementation steps and formula substitution: Baseline calibration and construction (S100): An automated drilling rig pushes an anchor bolt body equipped with a fiber Bragg grating sensor array into the borehole. Data is acquired via a drilling-while-drilling module, including the cross-sectional area of ​​the drill bit. =0.0012m2, at the hole depth Measured propulsion pressure at 1.5m =8kN, the mechanical power component converted from speed and torque is 12kW, and the advance speed is... =0.02m / s. The system substitutes this value into the formula and calculates the drilling specific energy at that depth in real time. An initial dynamic baseline was constructed, and the acousto-optic baseline was recorded after applying preload.

[0098] Regional integrated state solution (S400): After multi-pore collaborative detection and service-life ultrasonic active detection, the edge computing controller evaluates a section of voxel mesh on the right side of the roadway. The normalized transpore attenuation factor for this region is extracted. =0.8 (indicating numerous fractures in the surrounding rock), and the normalized interfacial decoupling index. =0.7 (indicating a risk of peeling at the anchorage interface).

[0099] Set fusion weights The system substitutes the values ​​into the fusion formula to calculate the regional comprehensive state index of the grid: ; Adaptive adjustment and closed-loop execution of support parameters: The system's preset safety threshold =0.6. Because (0.74) > 0.6, the system determines that the area is a high-risk weak geological zone.

[0100] Substitute the parameters into the dynamic compensation formula for grouting pressure (set the compensation coefficient). =0.4): Normalized excess ratio

[0101] Dynamic grouting pressure

[0102] Subsequently, the edge computing controller sends a grouting command of 2.28 MPa to the intelligent grouting pump, which performs pressurized closed-loop grouting on the weak surface to induce grout to enter the micro-cracks and complete self-correction.

[0103] Experimental verification and Figure 3 Comparison of effects: To verify the effectiveness of the adaptive correction control mechanism based on the state evaluation matrix of this invention, two 100-meter test areas with similar geological conditions were selected in the aforementioned 8304 return air trough for comparative experiments (the service monitoring cycle is 180 days), and the results were generated as follows: Figure 3 The chart shown is a comparative analysis of the evolution of the interface deadhesion index during service life.

[0104] exist Figure 3 In the overall layout, the legend in the upper left box is used to correspond to the specific technical solutions (i.e., traditional solutions and the system of this invention) represented by different line types and marking symbols, so that readers can quickly distinguish between the two verification data chains. Figure 3 The meanings of the coordinate system and the baseline are as follows: The horizontal axis (X-axis) and service time (days) represent the duration for which the anchor bolt support system has been installed and put into use downhole. The statistical period is set from the completion of installation to 180 days. This reflects the continuous test of the stability of the support system over time under the long-term rheological effects of ground stress.

[0105] Vertical axis (Y-axis), interface de-adhesion index The dimensionless physical index calculated in the embodiments ranges from 0 to 1. This index is used to quantitatively characterize the bonding state between the anchor body and the surrounding rock interface. The judgment criteria are: the closer the value is to 0, the tighter the interface bonding and the better the acoustic impedance matching; the closer the value is to or exceeds a certain threshold, the more serious the interface debonding and peeling, and the more obvious the grouting voids or failure risk.

[0106] Horizontal black dotted line: represents the safety threshold ( =0.6), this is the system-set support failure early warning red line. When the actual measurement When this boundary is crossed, it physically means that there has been a large-scale substantial debonding at the anchor bolt interface, the anchoring force has been greatly reduced, and the roadway has an extremely high risk of instability and collapse.

[0107] Combination Figure 3 The evolution of the data for the two test areas is as follows: Control group (traditional scheme): conventional anchor bolts were used, and a constant grouting pressure of 2.0MPa was applied throughout the entire section without any adaptive parameter adjustment.

[0108] exist Figure 3 In the figure, the data curve for the control group is represented by a solid black line with hollow circles. Under long-term rheological disturbances of in-situ stress, this curve shows a clear upward trend. This indicates that under constant grouting pressure, due to the inability to fill deep micro-cracks, the interfacial micro-cracks continue to expand with the passage of service time, leading to gradual separation of the grout from the surrounding rock. Specifically, after 30 days of service, The deadhesion index rapidly increased from an initial 0.22, leading to a continuous deterioration in its deadhesion index; in the later stages of service, Figure 3 The solid line in the middle broke through. =0.6 horizontal black dotted line (safety red line); by 180 days, It surged to 0.68, reaching a dangerous state of large-area substantial debonding failure at the interface.

[0109] Experimental group (system of this invention): Using this system, based on the aforementioned calculations... The index and multidimensional state assessment matrix were used to implement adaptive densification and dynamic pressure compensation (such as local pressure increase to 2.28~2.50MPa) correction of the anchor bolt measurement section by automated drilling rig and intelligent grouting pump.

[0110] exist Figure 3In the diagram, the data curve for this experimental group is represented by a black dashed line with hollow squares. This curve remained stable and low throughout the 180 days, consistently kept below the safety threshold (horizontal black dotted line). The data indicates that, due to the adaptive pressure grouting implemented during construction, the active correction mechanism of this invention effectively promoted the full penetration of the grout into the weak zone, significantly enhancing the interfacial coupling capacity. During its service life, even under long-term geostress disturbances, the IDI index only slowly increased from the initial 0.20 to 0.28, and the anchorage interface remained highly dense.

[0111] in conclusion: Experimental verification and Figure 3 The trend clearly shows that the multi-source parameter fusion and adaptive closed-loop control scheme in this invention effectively avoids the local debonding failure caused by the traditional one-size-fits-all constant grouting blind support. It has obvious anti-debonding and anti-failure effects, and greatly improves the safety and reliability of the support system in the entire life cycle of underground coal mines.

Claims

1. A fully life-cycle intelligent sensing anchor bolt support system for underground coal mines, characterized in that, include: An anchor bolt body, wherein a fiber Bragg grating sensor array and a piezoelectric ceramic transducer are integrated on the anchor bolt body; Construction machinery, including automated drilling rigs and intelligent grouting pumps; The edge computing controller is communicatively connected to the fiber Bragg grating sensor array, piezoelectric ceramic transducer, and automated drilling rig, and is used for: Extract the mechanical parameters during the construction period and the initial acoustic and optical signals of the anchor bolt body to establish a state baseline, which includes the drilling dynamics baseline and the initial acoustic and optical state baseline. The mechanical vibration generated by the automated drilling rig operation in adjacent holes is used as the sound source for detection. Combined with the drilling dynamics baseline, the cross-hole attenuation factor is calculated by receiving the cross-hole acoustic emission signal. The piezoelectric ceramic transducer is triggered to emit ultrasonic waves, which are received by the fiber Bragg grating sensor array. Using the initial acousto-optic state baseline as a reference, the interface decoupling index is calculated by combining the transpore attenuation factor to achieve state decoupling. The cross-hole attenuation factor and the interface debonding index are combined to evaluate the regional status, and the construction parameters are corrected accordingly and sent to the automated drilling rig for closed-loop control.

2. The intelligent sensing anchor bolt support system for the entire life cycle of coal mines according to claim 1, characterized in that: The fiber Bragg grating sensor array is embedded in a groove on the surface of the anchor body; The piezoelectric ceramic transducer is integrated into the end region of the anchor bolt body; The anchor bolt body is fixedly installed with a connector assembly at its tail. The connector assembly is equipped with a beam-expanding magnetic photoelectric composite interface, and internally encapsulates a miniature collimating lens group and a ring power supply contact, and integrates a miniature drive control circuit. The miniature collimating lens group is connected to the fiber Bragg grating sensor array, and the annular power supply contact is electrically connected to the miniature drive control circuit and the piezoelectric ceramic transducer.

3. The intelligent sensing anchor bolt support system for the entire life cycle of coal mines according to claim 2, characterized in that: The power head of the automated drilling rig is equipped with a drilling-while-drilling sensing module, a displacement sensor, and a magnetic follow-up connector. The magnetic follow-up connector is connected to the beam-expanding magnetic photoelectric composite interface. The automated drilling rig's onboard platform is equipped with a high-speed dynamic demodulator; The high-speed dynamic demodulator acquires the optical signals of the fiber Bragg grating sensor array through the magnetic servo connector; The miniature collimating lens group is used to amplify the single-mode beam in parallel divergence, and to construct an optical transmission channel that can tolerate a certain amount of dust particle obstruction.

4. The intelligent sensing anchor bolt support system for the entire life cycle of coal mines according to claim 3, characterized in that, The specific process for establishing the drilling dynamics baseline includes: By integrating the instantaneous advance depth recorded by the displacement sensor, the drilling rig push pressure, slewing angular velocity, slewing torque and instantaneous advance speed obtained by the drilling while sensing module are resampled and mapped to the same depth coordinate system for data alignment. Invalid records that are filtered out when the instantaneous advance speed is lower than the preset dead zone threshold or when the advance pressure drops sharply are removed. A cubic spline interpolation algorithm is used to process missing data segments to ensure the continuity of the spatial sequence. Substitute the valid data at the same depth coordinates into the drilling specific energy calculation formula to calculate the unit rock breaking energy consumption value at each depth, and store it in the form of a spatial sequence as the drilling dynamics baseline.

5. The intelligent sensing anchor bolt support system for the entire life cycle of coal mines according to claim 3, characterized in that, The specific process for establishing the initial acoustic-optical state baseline includes: The actual reflection center wavelength of the fiber Bragg grating sensor array after the initial preload is applied and temperature compensation is completed is obtained, and the initial static strain field is calculated by combining the factory calibration wavelength data. The high-frequency pulse excitation signal is sent to the piezoelectric ceramic transducer through the ring power supply contact to generate ultrasonic waves. The high-speed dynamic demodulator synchronously captures the dynamic wavelength drift signal and performs square integration within a set time window to obtain the energy spatial distribution parameters of the initial transmitted wave. The initial static strain field is bound to the energy spatial distribution parameter of the initial transmitted wave and encapsulated as the initial acousto-optic state baseline.

6. The intelligent sensing anchor bolt support system for the entire life cycle of coal mines according to claim 3, characterized in that, The processing and feature extraction of the transap-hole acoustic emission signal includes: The high-frequency fluctuation component of the instantaneous rotational torque of the automated drilling rig during rock breaking operations is extracted as a discrete mechanical prior reference template sequence; Envelope detection is performed on the optical wavelength drift sequence captured by the fiber Bragg grating sensor array, and downsampling interpolation is performed based on the mechanical basic sampling frequency to obtain a time-aligned optical wavelength drift sequence at the receiver. Within a delay time window defined based on the spatial linear propagation distance, a sliding cross-correlation operation is performed between the discrete mechanical prior reference template sequence and the receiver optical wavelength drift sequence; Peak search is performed on the cross-correlation function value sequence of the source and receiver signals, and the square amplitude of the maximum peak value is extracted as the effective transmitted wave energy parameter after enhancement.

7. The intelligent sensing anchor bolt support system for the entire life cycle of coal mines according to claim 6, characterized in that, The process of calculating the transaperture attenuation factor includes: Multiply the enhanced effective transmitted wave energy parameter by the square of the linear propagation distance between the rock-breaking source and the grating node to calculate the equivalent wavefront energy after geometric diffusion compensation. The relative transmission energy ratio is calculated by dividing the equivalent wavefront energy by the initial mechanical radiation energy reference value and multiplying it by the end-to-end uniform calibration constant. Based on the exponential decay model, a logarithmic transformation is performed on the relative transmission energy ratio to calculate the transaperture attenuation factor connecting the corresponding spatial paths.

8. The intelligent sensing anchor bolt support system for the entire life cycle of coal mines according to claim 3, characterized in that, The state decoupling process includes: The edge computing controller extracts the reflected wave sequence corresponding to the ultrasonic wave, and after DC bias removal and bandpass filtering, calculates the sum of squares of the amplitude of the sampling points, and divides it by the preset reference excitation energy constant to solve for the relative reflection energy ratio. Extract the reference value of the equivalent trans-aperture attenuation factor corresponding to the current monitoring node after spatial interpolation preprocessing; The measured relative reflection energy ratio and the equivalent trans-hole attenuation factor reference value are substituted into the nonlinear decoupling model to calculate the interface debinding index. The calculated value of the interface debonding index is the quotient of the relative reflection energy ratio divided by the acoustic impedance compensation term, wherein the acoustic impedance compensation term includes a negative exponential term of the equivalent transaperture attenuation factor reference value.

9. The intelligent sensing anchor bolt support system for the entire life cycle of underground coal mines according to claim 1, characterized in that, The process of assessing the state of the region includes: The target support area is discretized in three-dimensional space using a regular voxel mesh; Using a three-dimensional inverse distance weighted interpolation algorithm, the trans-hole attenuation factor and the interface deadhesion index in the local sensor coordinate system are aligned to the corresponding three-dimensional voxel units; Extreme value normalization was performed on the transpore attenuation factor and the interface debonding index within the same three-dimensional voxel unit, and then linearly added together with the set fusion weight coefficient to calculate the regional comprehensive state index characterizing the health of local support. The regional comprehensive state indices of all three-dimensional voxel units are tensor-stitched according to spatial topological relationships to construct a regional state evaluation matrix.

10. The intelligent sensing anchor bolt support system for the entire life cycle of coal mines according to claim 9, characterized in that, The process of correcting construction parameters and implementing closed-loop control includes: Voxel units whose regional comprehensive state index exceeds the preset safety critical threshold are selected from the regional state assessment matrix and connected and delineated as weak geological zones using a spatial density clustering algorithm. The normalized over-limit ratio is calculated based on the difference between the arithmetic mean of the comprehensive state index of the voxel unit region within the weak geological zone and the preset safety critical threshold. The normalized excess ratio is multiplied by the preset grouting pressure compensation coefficient and a constant 1 is added. Then, it is multiplied by the preset standard basic grouting pressure constant to calculate the target dynamic grouting pressure. The standard industrial control message containing the three-dimensional coordinates of the reinforced borehole and the target dynamic grouting pressure is sent to the automated drilling rig and intelligent grouting pump to perform targeted densified drilling and dynamic pressure boosting grouting operations.

Images