A drilling and blasting method high ground stress rock burst intelligent early warning prevention and control system

Abstract

The present application relates to the technical field of tunnel engineering disaster monitoring and prevention, and discloses a drilling and blasting method high ground stress rock burst intelligent early warning and prevention system, which comprises an early warning system, a rock drilling jumbo, a microseismic system and a high-pressure water injection device. The present application fuses the rock drilling jumbo operation mechanical parameter and the microseismic system tunnel environment noise data to form a multi-modal sensing data sequence, inputs an hidden Markov model to identify the construction stage and switches the operation branch; calculates the drilling specific energy to update the elastic wave propagation speed three-dimensional matrix, combines the blasting damage wave velocity reduction matrix to locate the microseismic event and generates the dangerous aggregation volume; calculates the spatial intersection volume of the dangerous aggregation volume and the actual pressure relief wave propagation volume to obtain the dangerous area coverage rate and issues the control instruction, calculates the water injection pressure increment by using the residual error of the actually occurring drilling specific energy and the initial estimated drilling specific energy, calls the penalty factor to perform the iterative correction operation on the spatial interpolation algorithm to complete the bottom logic closed loop self-correction.

Description

Technical Field

[0001] This invention relates to the field of disaster monitoring and prevention technology in tunnel engineering, specifically to an intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method. Background Technology

[0002] Existing monitoring systems for deep tunnel drilling and blasting construction independently collect drilling machinery parameters and microseismic waveform signals. These systems lack a multimodal sensor data fusion mechanism, failing to automatically identify tunnel construction stages based on machinery parameters and tunnel environmental noise data, resulting in asynchronous switching of the system's underlying computational processing logic. When processing microseismic waveform signals, existing systems rely on a fixed elastic wave propagation velocity matrix, failing to dynamically update the wave velocity matrix node data by incorporating uniaxial compressive strength correlated with drilling specific energy, and neglecting to account for wave velocity reduction caused by blasting operations. This static wave velocity model causes the three-dimensional spatial location of microseismic events to deviate from the true coordinates, reducing the accuracy of determining the three-dimensional spatial coordinates of microseismic event clusters.

[0003] Current early warning and control measures lack a spatial quantitative assessment of the results of water injection and pressure relief operations. After high-pressure water injection, existing systems cannot establish a spatial intersection calculation model between the actual pressure relief volume and the dangerous accumulation volume, and therefore cannot quantitatively output the coverage rate of the dangerous area. Existing systems lack a mechanism to reverse-adjust the water injection pressure using the residual between the actual drilling energy and the initially estimated drilling energy, and cannot invoke the underlying penalty factor to iteratively correct the spatial interpolation algorithm. This results in a unidirectional output of on-site control parameters, leaving blind spots for rockburst hazards.

[0004] Therefore, this invention proposes an intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method to address the shortcomings of existing technologies. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method. This system solves the problems of existing monitoring systems, such as the inability to integrate multimodal sensor data to identify the construction stage, resulting in the inability of the underlying computational logic to automatically switch; the reliance on static wave velocity matrices leading to deviations in the three-dimensional spatial positioning of microseismic events from their physical coordinates; and the lack of quantitative assessment of pressure relief space and specific energy residual feedback mechanisms, resulting in the inability of prevention and control parameters to self-correct in a closed loop.

[0006] To achieve the above objectives, the present invention provides the following technical solution: This invention provides an intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method, comprising an early warning system, a drilling rig, a micro-vibration system, and a high-pressure water injection device. The early warning system establishes bidirectional data communication links with the drilling rig, the micro-vibration system, and the high-pressure water injection device. The drilling rig is equipped with a data bus and a spatial positioning unit. The drilling rig transmits operational machinery parameters to the early warning system via the data bus, and uses the spatial positioning unit to acquire the three-dimensional spatial coordinates of the drill bit and send these coordinates back to the early warning system. The micro-vibration system is equipped with a detector array. The micro-vibration system uses the detector array to collect tunnel environmental noise data and micro-vibration event waveform signals, and then transmits these data to the early warning system. The high-pressure water injection equipment is equipped with a frequency converter and a pressure and flow sensor module. The high-pressure water injection equipment receives the three-dimensional coordinate suggestion command for hole filling and water injection and the secondary water injection control command issued by the early warning system through the frequency converter. The high-pressure water injection equipment uses the pressure and flow sensor module to monitor the pipeline pressure drop slope and the cumulative water injection volume in real time and feeds the pipeline pressure drop slope and the cumulative water injection volume back to the early warning system.

[0007] The early warning system is equipped with a calculation module and a storage module. The storage module pre-stores a hidden Markov model, a rock mechanics mapping table, and a three-dimensional matrix of elastic wave propagation velocity. The early warning system uses the calculation module to read the operating machinery parameters sent by the drilling rig, the tunnel environmental noise data and microseismic event waveform signals transmitted by the microseismic system, and the data fed back by the high-pressure water injection equipment. The early warning system inputs the operating machinery parameters and tunnel environmental noise data into the hidden Markov model, and the hidden Markov model outputs the real-time stage of tunnel construction.

[0008] When tunnel construction is in the drilling stage, the early warning system extracts the parameters of the operating machinery and calculates the drilling specific energy. The system inputs this drilling specific energy into a rock mechanics mapping table, which outputs a uniaxial compressive strength distribution sequence. The system then updates the three-dimensional matrix of elastic wave propagation velocity using this sequence. When tunnel construction is in the post-blasting redistribution stage, the system calls the updated three-dimensional matrix of elastic wave propagation velocity to perform three-dimensional spatial positioning calculations on the microseismic event waveform signals. The system generates a high-stress rockburst early warning signal and marks the microseismic event cluster area within the three-dimensional spatial model, defining it as a dangerous accumulation volume. The system monitors data from the high-pressure water injection equipment and captures high-frequency hydraulic fracturing acoustic emission events accompanying the water injection operation. It calculates the envelope three-dimensional space formed by these events and defines it as the actual pressure relief volume. The system then uses the dangerous accumulation volume and the actual pressure relief volume to calculate and output the dangerous area coverage rate. Based on this dangerous area coverage rate, the system sends instructions to the high-pressure water injection equipment and extracts the intersection region. The early warning system continuously monitors the real-time stage changes of tunnel construction. When the three-dimensional spatial coordinates of the drilling rig fall into the intersection area, the early warning system re-collects the operating machinery parameters and generates the actual drilling energy. The early warning system extracts the initial estimated drilling energy and calculates the drilling energy residual. Based on the drilling energy residual, the early warning system triggers the bottom-level linkage control operation. The early warning system reversely adjusts the rock mechanics mapping table and updates the three-dimensional matrix of elastic wave propagation velocity to complete the bottom-level logic closed-loop self-correction.

[0009] The early warning system utilizes the tunnel reference coordinates, dead reckoning data, and absolute encoder angle data output by the spatial positioning unit to track and calculate the dynamic spatial attitude between the global coordinate system based on the tunnel's absolute geodetic coordinates and the local coordinate system of the drilling rig. The system performs frame processing on the tunnel environmental noise data according to a preset time sliding window scale and extracts the frequency domain energy spectrum feature vector. It also performs time-axis alignment operations based on timestamps on the operating machinery parameters, and weightedly concatenates the dimensionality-reduced frequency domain features with the operating machinery parameters to form a joint feature matrix. The system generates a multimodal sensor data sequence based on discretized finite-state dictionary data and inputs this sequence into a Hidden Markov Model (HMM). The system calculates the maximum joint probability value of each state within the hidden state set corresponding to the multimodal sensor data sequence generated by the HMM at the current time node. Finally, the system extracts the state associated with the maximum joint probability value and defines it as the real-time stage of tunnel construction.

[0010] The early warning system combines drilling energy and the three-dimensional spatial coordinates of the drill string to generate spatial energy data nodes, and inputs them into a rock mechanics mapping table to output a uniaxial compressive strength distribution sequence. Using the three-dimensional spatial coordinates and uniaxial compressive strength values ​​recorded within the uniaxial compressive strength distribution sequence, the early warning system dynamically outputs the reference velocity for elastic wave propagation in the rock medium and defines it as the elastic wave propagation velocity value, synchronously replacing the original velocity data stored within the matrix nodes to complete the update process.

[0011] The early warning system uses a blast damage wave velocity reduction coefficient to perform a velocity reduction correction on the three-dimensional elastic wave propagation velocity matrix, generating a blast damage attenuation wave velocity matrix. The system then uses this attenuation matrix to perform three-dimensional spatial positioning calculations on the microseismic event waveform signal, calculating and outputting the three-dimensional spatial coordinates of the microseismic event. Based on these coordinates and the released energy, the system calculates the microseismic spatial energy density. Finally, the system uses a spatial density clustering algorithm to perform spatial point cloud aggregation calculations on the three-dimensional spatial coordinates of the microseismic event, generating a hazardous accumulation volume.

[0012] The high-pressure water injection equipment executes water injection operations based on high-stress rockburst early warning signals, inducing fracture propagation within the rock mass. The early warning system utilizes the transient wave velocity matrix of water-induced fracturing to perform three-dimensional spatial positioning calculations on the high-frequency hydraulic fracturing acoustic emission waveform sequence, generating an acoustic emission spatial coordinate point cloud. The early warning system calculates the spatial volume parameters within the closed polyhedral structure and defines them as the actual pressure relief swept volume. The early warning system calculates the overlap between the actual pressure relief swept volume and the hazardous accumulation volume to generate a spatial intersection volume. The early warning system performs a division operation on the original spatial volumes corresponding to the spatial intersection volume and the hazardous accumulation volume to obtain the hazardous area coverage. Within the tunnel's three-dimensional spatial model, the early warning system subtracts the three-dimensional intersection boundary coordinates from the three-dimensional spatial boundary coordinates to obtain the uncovered hazardous areas. The early warning system calls a clustering algorithm to segment the uncovered hazardous point cloud into regional clusters and calculates target point coordinates to generate borehole filling spatial trajectory parameters. These parameters are then sent to the drilling rig to perform positioning and borehole filling operations. The early warning system reads the three-dimensional spatial coordinates of the drilling tool for the new round and extracts the initial estimated drilling energy. It then performs a subtraction operation between the actual drilling energy and the initial estimated drilling energy to obtain the drilling energy residual. Using this residual, the system calculates the water injection pressure increment and adjusts the output water injection pressure. Finally, it calls a penalty factor to iteratively correct the three-dimensional spatial interpolation algorithm, completing a logical closed-loop self-correction process.

[0013] This invention provides an intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method. It has the following beneficial effects: 1. This invention combines the operating mechanical parameters of the rock drilling rig and the tunnel environmental noise data of the microseismic system to form a multimodal sensing data sequence, and inputs it into a hidden Markov model for joint probability calculation. The system identifies the real-time stage of tunnel construction, and automatically triggers the corresponding drilling energy calculation logic and microseismic signal analysis logic based on the transition results of the drilling stage and the post-blasting redistribution stage, thus completing the automatic switching of the low-level calculation branch for early warning and prevention.

[0014] 2. This invention calculates the drilling specific energy and uses a rock mechanics mapping table to output a uniaxial compressive strength distribution sequence, updating the nodal velocity data within the three-dimensional matrix of elastic wave propagation velocity. The system combines the blast damage wave velocity reduction coefficient to generate a blast damage attenuation wave velocity matrix, and uses this matrix to perform three-dimensional spatial positioning calculations on the microseismic event waveform signal, improving the consistency between the three-dimensional spatial coordinates of the microseismic event cluster area and the on-site spatial conditions.

[0015] 3. This invention obtains the spatial intersection volume by calculating the overlap between the dangerous accumulation volume and the actual pressure relief spillover volume, and then uses the spatial intersection volume and the dangerous accumulation volume to calculate the dangerous area coverage rate. Based on the dangerous area coverage rate determination result, the system issues secondary pressure relief or hole drilling commands. Simultaneously, it calculates the water injection pressure increment using the residual between the actual drilling specific energy and the initially estimated drilling specific energy, and calls the penalty factor to perform iterative correction operations on the spatial interpolation algorithm, completing the underlying logic closed-loop self-correction of the control parameters. Attached Figure Description

[0016] Figure 1 This is a system block diagram of the present invention; Figure 2 This is a comparison chart of actual drilling energy under different pressure relief control strategies in a specific application embodiment of the present invention. Detailed Implementation

[0017] 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.

[0018] See attached document Figure 1This invention provides an intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method, comprising an early warning system, a drilling rig, a micro-vibration system, and a high-pressure water injection device. The early warning system establishes bidirectional data communication links with the drilling rig, the micro-vibration system, and the high-pressure water injection device. The drilling rig is equipped with a data bus and a spatial positioning unit. The drilling rig transmits the operating machinery parameters to the early warning system through the data bus. The drilling rig uses the spatial positioning unit to obtain the three-dimensional spatial coordinates of the drill bit and sends the three-dimensional spatial coordinates of the drill bit to the early warning system.

[0019] Due to the lack of satellite positioning signals in the deep environment of the tunnel, the spatial positioning unit is composed of a laser total station target prism pre-installed on the tunnel wall, an inertial measurement unit (IMU) arranged on the body of the rock drilling rig, and an absolute encoder at the joint of the drill arm. The early warning system obtains the reference coordinate points of the tunnel from the calibration output of the laser total station in advance, and uses the real-time attitude data of the inertial measurement unit to perform dead reckoning.

[0020] The early warning system uses the cavern reference coordinates, dead reckoning data, and absolute encoder angle data output by the rock drilling rig's spatial positioning unit to track and calculate the dynamic spatial attitude between the global coordinate system based on the absolute geodetic coordinates of the tunnel and the local coordinate system of the rock drilling rig in real time. The early warning system continuously updates the affine transformation matrix between the two and calls the updated affine transformation matrix to perform coordinate system transformation operations on the three-dimensional spatial coordinates of the drill bit, aligning the three-dimensional spatial coordinates of the drill bit to the global coordinate system.

[0021] The microseismic system is equipped with a detector array, which collects tunnel environmental noise data and microseismic event waveform signals. To accommodate both low-frequency microseismic signals and high-frequency acoustic emission signals, the detector array combines low-frequency microseismic sensors with broadband acoustic emission sensors to ensure that the hardware monitoring frequency band covers the high-frequency hydraulic fracturing acoustic signals generated during fracture propagation. The microseismic system transmits tunnel environmental noise data and microseismic event waveform signals to the early warning system in real time. The high-pressure water injection equipment is equipped with a frequency converter and a pressure and flow sensor module. The high-pressure water injection equipment receives the three-dimensional coordinate suggestion command for borehole water injection and the secondary water injection control command issued by the early warning system through the frequency converter. The high-pressure water injection equipment uses the pressure and flow sensor module to monitor the pipeline pressure drop slope and cumulative water injection volume in real time and feeds these parameters back to the early warning system.

[0022] The early warning system is equipped with a computing module and a storage module. The storage module pre-stores a hidden Markov model, a rock mechanics mapping table, and a three-dimensional matrix of elastic wave propagation velocity. The early warning system uses the computing module to read the operating machinery parameters sent by the rock drilling rig, the microseismic event waveform signals transmitted by the microseismic system, and the data fed back by the high-pressure water injection equipment. The early warning system performs multi-source data fusion calculation and three-dimensional spatial matching calculation. Finally, the early warning system outputs control commands to the high-pressure water injection equipment to form a closed-loop data communication link.

[0023] The early warning system receives the operating machinery parameters sent by the rock drilling rig and the tunnel environmental noise data and micro-seismic event waveform signals transmitted by the micro-seismic system. The early warning system inputs the operating machinery parameters and tunnel environmental noise data into the hidden Markov model, and the hidden Markov model outputs the real-time stage of tunnel construction.

[0024] When the tunnel construction is in the drilling stage, the early warning system extracts the parameters of the operating machinery and calculates the drilling specific energy. The early warning system inputs the drilling specific energy into the rock mechanics mapping table, and the rock mechanics mapping table outputs the uniaxial compressive strength distribution sequence. The early warning system uses the uniaxial compressive strength distribution sequence to update the three-dimensional matrix of elastic wave propagation velocity.

[0025] When the tunnel construction is in the post-blasting redistribution stage, the early warning system calls the updated elastic wave propagation velocity three-dimensional matrix to perform three-dimensional spatial positioning calculation on the microseismic event waveform signal. The early warning system generates a high ground stress rockburst early warning signal. The early warning system marks the microseismic event cluster area in the three-dimensional spatial model and defines the microseismic event cluster area as the dangerous accumulation volume.

[0026] The early warning system monitors the data fed back by the high-pressure water injection equipment and captures the high-frequency effective hydraulic fracturing acoustic emission events generated during the water injection operation. The early warning system calculates the envelope three-dimensional space formed by the high-frequency effective hydraulic fracturing acoustic emission events and defines the envelope three-dimensional space as the actual pressure relief wave volume. The early warning system calls the dangerous accumulation volume and the actual pressure relief wave volume to calculate and output the dangerous area coverage rate. The early warning system issues instructions to the high-pressure water injection equipment based on the dangerous area coverage rate and extracts the intersection area.

[0027] The early warning system continuously monitors the real-time stage changes of tunnel construction. When the three-dimensional spatial coordinates of the drilling rig fall into the intersection area, the early warning system re-collects the operating machinery parameters and generates the actual drilling energy. The early warning system extracts the initial estimated drilling energy and calculates the drilling energy residual. Based on the drilling energy residual, the early warning system triggers the bottom-level linkage control operation. The early warning system reversely adjusts the rock mechanics mapping table and updates the three-dimensional matrix of elastic wave propagation velocity to complete the bottom-level logic closed-loop self-correction.

[0028] The early warning system receives the operating machinery parameters sent by the rock drilling rig, and receives the tunnel environmental noise data collected and transmitted by the microseismic system. The operating machinery parameters include the drilling rig thrust, drilling rig rotation speed, drilling rig rotation torque and drilling rig penetration rate. The microseismic system uses a detector array to continuously sample and convert the acoustic waveforms inside the tunnel to generate tunnel environmental noise data.

[0029] The early warning system periodically broadcasts Precise Time Synchronization Protocol (PTP) messages to the drilling rig and microseismic system via a two-way data communication link. The early warning system compensates for the drift of local quartz crystal oscillators in each underlying hardware and network transmission delay, unifying the global absolute time reference when each independent device generates data frames. The early warning system sets up a circular time buffer queue at the receiving end. The early warning system receives data from the data bus and detector array, extracts the timestamp records corresponding to the working machinery parameters and the timestamp records corresponding to the tunnel environmental noise data, and performs time-series soft alignment operation based on timestamps in the circular time buffer queue using a sliding time window mechanism. Subsequently, the early warning system first performs filtering and noise reduction processing and amplitude normalization calculation on the high sampling rate tunnel environmental noise data.

[0030] The early warning system performs frame processing on the tunnel environmental noise data after amplitude normalization according to a preset time sliding window scale, and extracts the frequency domain energy spectrum feature vector within each time sliding window, thereby reducing the dimensionality of the high-frequency noise data into a frequency domain feature sequence with low temporal resolution. The early warning system performs filtering and noise reduction and amplitude normalization on the operating machinery parameters with low sampling rate. Subsequently, the early warning system performs time axis alignment operation based on timestamps on the dimensionality-reduced frequency domain feature sequence and the operating machinery parameters, using the preset time sliding window scale as a benchmark.

[0031] The early warning system detects the continuity of the operating machinery parameters after time alignment. When it identifies that the operating machinery parameters are missing due to shutdown, the early warning system inserts a zero vector placeholder containing the shutdown status identifier into the missing timestamp node to complete the data, so as to indicate that the machinery is in a non-rock breaking operation state. The early warning system first calls the principal component analysis (PCA) algorithm to perform preliminary dimensionality reduction on the high-dimensional frequency domain energy spectrum feature vector and extracts the principal components with a cumulative contribution rate greater than a set threshold.

[0032] Subsequently, the early warning system weighted and concatenated the dimensionality-reduced frequency domain features with the operating machinery parameters to form a joint feature matrix of equal scale. During the weighted concatenation process, the early warning system scans the operating machinery parameters for zero vector placeholders containing shutdown status identifiers. For timestamp nodes containing zero vector placeholders, the early warning system dynamically sets the concatenation weight of the corresponding machinery parameters to zero and triggers a feature masking mechanism to shield the interference of extremely low variance in invalid data caused by equipment shutdown. The early warning system calls the Gaussian mixture clustering algorithm to perform joint dimensionality reduction and spatial mapping operations on the joint feature matrix. The early warning system calls the K-bins binning algorithm to establish multi-level threshold intervals. The early warning system uses the multi-level threshold intervals to discretize and map the continuous parameters after dimensionality reduction and mapping into finite state dictionary data. The early warning system generates multimodal sensing data sequences based on the finite state dictionary data.

[0033] The early warning system establishes a hidden Markov model, within which a set of hidden states is defined. This set of hidden states consists of the drilling stage, the charging stage, the post-blasting redistribution stage, the muck removal stage, and the support stage. The early warning system maps the operational actions occurring during the tunnel construction cycle to the drilling stage, the charging stage, the post-blasting redistribution stage, the muck removal stage, and the support stage.

[0034] The early warning system sets up a state transition probability matrix, an observation probability matrix, and an initial state probability vector within the Hidden Markov Model. It extracts historical tunnel construction stage records and generates an initial state probability vector. The system also extracts the temporal irreversibility characteristics of standard tunnel construction procedures, assigning a non-zero minimum penalty constant to the probability of impossible cross-stage jumps between the drilling stage, charging stage, post-blasting redistribution stage, muck removal stage, and support stage to prevent the algorithm from deadlocking due to abnormal jump conditions. A procedure transition logic with prior directed graph constraints is established, and the system sets the state transition probability matrix based on this logic. Furthermore, the system establishes frequency domain feature prior constraint rules based on the physical acoustic characteristics of tunnel construction. Specifically, the acoustic characteristics of the muck removal stage are mapped to a low-frequency broadband impact energy spectrum, the acoustic characteristics of the shotcrete operation in the support stage are mapped to a high-frequency narrowband continuous energy spectrum, and the acoustic characteristics of the charging stage are mapped to a background silent energy spectrum without peaks. The acoustic characteristics of the drilling stage are mapped to the medium-to-high frequency mechanical rock-breaking energy spectrum with periodic harmonic components, and the acoustic characteristics of the initial moment of the redistribution stage after blasting are mapped to the transient full-band impact energy spectrum with extremely high amplitude.

[0035] The early warning system combines frequency domain feature prior constraint rules to establish mapping rules between multimodal sensing data sequences and implicit state sets, and generates an observation probability matrix.

[0036] Specifically, during the offline phase, the early warning system extracts multimodal sensor data sequences of tunnel construction with manually marked historical stages of the process. The early warning system calls the expectation-maximization (EM) algorithm to estimate the parameters of the hidden Markov model. After multiple iterations until the likelihood function converges, the system trains and generates accurate initial baseline values ​​for the state transition probability matrix and the observation probability matrix.

[0037] The early warning system uses multimodal sensor data sequences to replace the observation variables set inside the hidden Markov model. The early warning system combines the state transition probability matrix, the observation probability matrix, and the initial state probability vector to complete the parameter initialization definition operation of the hidden Markov model.

[0038] The early warning system inputs multimodal sensor data sequences into a hidden Markov model. The early warning system uses the Viterbi decoding algorithm to perform probability calculations on the multimodal sensor data sequences. Within the hidden Markov model, the early warning system simultaneously calls the state transition probability matrix, the observation probability matrix, and the initial state probability vector.

[0039] The early warning system calculates the maximum joint probability value of each state in the hidden state set corresponding to the multimodal sensor data sequence generated by the hidden Markov model at the current time node. The early warning system performs numerical sorting and comparison operations on the maximum joint probability values ​​of each state, and extracts the state associated with the maximum joint probability value with the largest value in the numerical sorting and comparison operation.

[0040] The early warning system internally sets a confidence threshold for stage determination. It extracts the largest and second-largest joint probability values ​​and performs a subtraction operation to obtain the probability difference. This probability difference is then compared to the stage determination confidence threshold. When the probability difference is greater than the threshold, the system defines the state associated with the largest joint probability value as the real-time stage of tunnel construction. When the probability difference is less than or equal to the threshold, the system determines that it is currently in a transitional ambiguity period and suspends the determination, maintaining the tunnel construction real-time stage from the previous time point. If the system is in an initial startup state and there is no previous time point, it suspends the tunnel construction real-time stage by default until the post-blasting redistribution stage. The system also synchronously monitors the full-load status of the underlying hardware of the microseismic system's detector array. When the system detects tunnel environmental noise... When the amplitude of the data continuously exceeds the upper limit of the hardware range and waveform clipping occurs, the early warning system determines that the tunnel blasting has occurred. The early warning system triggers a transient hard interrupt bypass mechanism, forcibly suspends the current decoding process of the Hidden Markov Model, and forcibly overwrites the output of the real-time tunnel construction phase with the initial time node of the redistribution phase after blasting, thereby overcoming the deadlock caused by the extreme energy of the blasting moment on the Markov probability model. The early warning system continuously updates and outputs the real-time tunnel construction phase in time series, and triggers and controls the underlying related computational logic branches according to the real-time tunnel construction phase.

[0041] When the Hidden Markov Model outputs that the real-time stage of tunnel construction is the drilling stage, the early warning system initiates the logic for extracting operating machinery parameters. The early warning system reads the sensor value sequence sent by the drilling rig via the data bus, and internally sets lower thresholds for thrust and rotational speed.

[0042] The early warning system compares the thrust value recorded in the sensor data sequence with the lower limit threshold of thrust, and the rotational speed value recorded in the sensor data sequence with the lower limit threshold of rotational speed. When both the thrust value and the rotational speed value are greater than or equal to the lower limit threshold of thrust and rotational speed, the early warning system extracts the operating machinery parameters from the sensor data sequence.

[0043] The operating machinery parameters consist of drilling rig thrust, drilling rig rotation speed, drilling rig rotation torque, and drilling rig penetration rate. The early warning system stores the drilling rig thrust, drilling rig rotation speed, drilling rig rotation torque, and drilling rig penetration rate in the storage module according to the timestamp sequence, waiting to be called for calculation.

[0044] The early warning system reads the operating machinery parameters and borehole cross-sectional area stored in the storage module, and extracts the drilling rig thrust, drilling rig rotation speed, drilling rig rotation torque and drilling rig penetration rate from the operating machinery parameters.

[0045] The early warning system performs a division operation on the drilling rig thrust and borehole cross-sectional area to obtain the axial rock breaking energy. The early warning system performs a multiplication operation on the number 2, the constant of pi, the drilling rig rotation speed and the drilling rig rotation torque to obtain the rotational power. The early warning system divides the rotational power by the product of the borehole cross-sectional area and the drilling rig penetration rate to obtain the tangential rock breaking energy. The early warning system adds the axial rock breaking energy and the tangential rock breaking energy to obtain the drilling energy. The early warning system calculates the drilling energy according to the following formula.

[0046] ; In the formula, Represents drilling energy; Represents the drilling rig's propulsion force; Represents the cross-sectional area of ​​the borehole; Represents the rotational speed of the drilling rig; Represents the drilling rig's rotational torque; Represents the drilling rig's penetration rate; It represents the constant value of pi.

[0047] The early warning system establishes a rock mechanics mapping table, which establishes a mapping relationship between drilling specific energy and uniaxial compressive strength. The rock mechanics mapping table is pre-constructed using a polynomial fitting algorithm based on uniaxial compression test data of rock mass from historical tunnel exploration. The table contains pre-set conversion and correlation formulas composed of nonlinear empirical coefficients. The early warning system acquires the three-dimensional spatial coordinates of the drilling tool sent by the drilling rig using its spatial positioning unit. The system combines the drilling specific energy and the three-dimensional spatial coordinates of the drilling tool to generate spatial specific energy data nodes.

[0048] The early warning system arranges and aggregates spatial energy density data nodes into a continuous data stream according to timestamp order, and then inputs this continuous data stream into a rock mechanics mapping table. The rock mechanics mapping table performs lookup interpolation operations on the values ​​in the continuous data stream, and outputs a uniaxial compressive strength distribution sequence based on the lookup interpolation results.

[0049] The early warning system reads the three-dimensional matrix of elastic wave propagation velocity and the uniaxial compressive strength distribution sequence stored in the storage module. The early warning system extracts the three-dimensional spatial coordinates and uniaxial compressive strength values ​​recorded in the uniaxial compressive strength distribution sequence. The early warning system reads rock mass integrity coefficients pre-entered from historical geological exploration reports from the storage module, and also reads pre-imported tunnel advanced geological prediction reflected wave migration image data from the storage module. The system uses the reflected wave migration image data to extract the distribution range of fault fracture zones ahead of the tunnel face. It then uses this fault fracture zone distribution range to perform local attenuation assignment correction on the rock mass integrity coefficients pre-entered from historical geological exploration reports, generating dynamically corrected integrity coefficients. The system uses a global coordinate system based on the tunnel's absolute geodetic coordinates to perform three-dimensional spatial discretization and gridding of the dynamically corrected integrity coefficients. Using the three-dimensional spatial coordinates recorded within the uniaxial compressive strength distribution sequence, the system addresses and extracts matching local rock mass integrity coefficients within the discrete grid. The system internally pre-sets a wave velocity conversion function based on an empirical power function model. It synchronously inputs the uniaxial compressive strength values ​​and the extracted local rock mass integrity coefficients into the wave velocity conversion function. After calculation and processing, the wave velocity conversion function dynamically outputs the reference velocity for elastic wave propagation in the rock mass medium and defines it as the elastic wave propagation velocity value.

[0050] The early warning system uses three-dimensional spatial coordinates to perform coordinate matching and addressing operations within the three-dimensional matrix of elastic wave propagation velocity. The early warning system locates the matrix node corresponding to the three-dimensional spatial coordinates and replaces the original velocity data stored inside the matrix node with the elastic wave propagation velocity value.

[0051] The early warning system traverses the single-axis compressive strength distribution sequence and repeatedly executes coordinate matching addressing and data replacement operations to complete the replacement of the original velocity data of the three-dimensional spatial nodes of the borehole trajectory. Subsequently, to avoid blocking of the system's main control thread, the early warning system calls the Kriging three-dimensional spatial interpolation algorithm in the background asynchronous computing queue. The early warning system extracts the geological structure dip angle revealed by the tunnel's advanced geological prediction and the medium impedance variation curve along the tunnel axis (Z-axis). The early warning system uses the geological structure dip angle as the prior directional constraint condition of the anisotropic variability function and the medium impedance variation curve as the co-occurrence constraint. The Rikin interpolation incorporates a priori trend constraint term for the tunnel axial depth and sets an upper limit threshold for the extrapolation of the discrete variance of wave velocity to prevent divergence of predicted extrema. The early warning system uses the replaced three-dimensional spatial nodes of the borehole trajectory as reference control points and performs extrapolation assignment on the unexcavated three-dimensional unknown spatial nodes located in front of the tunnel face within the three-dimensional matrix of elastic wave propagation velocity to obtain interpolated wave velocity data. The early warning system uses the interpolated wave velocity data to update the three-dimensional unknown spatial nodes, completes the global update process of the three-dimensional matrix of elastic wave propagation velocity, and stores the updated three-dimensional matrix of elastic wave propagation velocity in the storage module.

[0052] When the hidden Markov model outputs the tunnel construction real-time stage to the post-blasting redistribution stage, the early warning system initiates the microseismic signal analysis logic. The early warning system receives the microseismic event waveform signals transmitted by the microseismic system, performs a jump point picking operation on the microseismic event waveform signals, and obtains the time parameters of the arrival time of the microseismic event waveform signals at the detector array inside the microseismic system.

[0053] The early warning system reads the updated three-dimensional matrix of elastic wave propagation velocity from the storage module. It extracts the single-hole charge amount from the tunnel blasting design parameters and calculates the depth of the loosened rock mass behind the tunnel face and the blasting damage velocity reduction coefficient based on a preset empirical formula for blasting damage attenuation. Using the blasting damage velocity reduction coefficient, the system performs velocity reduction correction on nodes within the updated three-dimensional matrix of elastic wave propagation velocity located within the loosened rock mass, generating a blasting damage attenuation velocity matrix. The system combines the time parameters of the microseismic event waveform signal with the blasting damage attenuation velocity matrix to generate a positioning input data stream. Using the blasting damage attenuation velocity matrix (replacing the updated three-dimensional matrix of elastic wave propagation velocity), the system performs three-dimensional spatial positioning calculations on the microseismic event waveform signal: the system calls the Fast Mover Method (FMM) or a three-dimensional ray tracing algorithm to perform forward modeling calculations within the updated three-dimensional matrix of elastic wave propagation velocity to obtain the theoretical travel time.

[0054] Subsequently, the early warning system verifies the number of geophone channels whose actual take-off point travel time has been extracted. When the number of geophone channels whose actual take-off point travel time has been extracted is greater than or equal to four, the early warning system triggers the inversion calculation logic branch. The early warning system constructs an objective function using the sum of the squared residuals between the theoretical travel time and the actual take-off point travel time of the microseismic event waveform signal, and calls the Geiger positioning algorithm for nonlinear iterative inversion until the objective function meets the convergence condition. Finally, the early warning system calculates and outputs the three-dimensional spatial coordinates of the microseismic event. The early warning system stores the three-dimensional spatial coordinates into the storage module in the order of occurrence time.

[0055] The early warning system performs integral calculations on the waveform signals of microseismic events and extracts the energy released by the microseismic events.

[0056] The early warning system reads the three-dimensional spatial coordinates stored in the storage module and iterates through all associated microseismic release energies within the three-dimensional spatial coordinate sequence. The system pre-defines three-dimensional mesh cells within the tunnel's three-dimensional spatial model and maps the three-dimensional spatial coordinates and microseismic release energies to the corresponding mesh cells. The system calculates the total accumulated microseismic release energy within each individual mesh cell and divides this total by the spatial volume of the mesh cell to obtain the microseismic spatial energy density. The system pre-defines a critical energy density threshold within the storage module. During non-operational quiet periods (such as the loading phase), the system extracts the background baseline mean of the microseismic spatial energy density and sets this threshold to the background baseline mean plus three standard deviations to adapt to the current background stress environment of the tunnel. The system compares the microseismic spatial energy density with the critical energy density threshold. When the microseismic spatial energy density exceeds the critical energy density threshold, the system determines that a localized high ground stress accumulation state has formed and triggers a hazard marker logic branch. When the spatial energy density of microseismic events exceeds the critical energy density threshold, the early warning system generates a high-stress rockburst early warning signal. The early warning system extracts the three-dimensional spatial coordinates of microseismic events associated with the three-dimensional grid cells that exceed the critical energy density threshold. The early warning system calls the spatial density clustering algorithm to perform spatial point cloud aggregation operations on the three-dimensional spatial coordinates of microseismic events, and calculates and generates microseismic event cluster areas.

[0057] The early warning system establishes a three-dimensional spatial model of the tunnel. The system imports the microseismic event cluster area into the tunnel's three-dimensional spatial model to perform boundary extraction and coordinate mapping operations. The system generates a closed geometric envelope contour within the tunnel's three-dimensional spatial model. The system extracts the three-dimensional volume parameters corresponding to the closed geometric envelope contour. The system defines the combination of the closed geometric envelope contour and the three-dimensional volume parameters as the hazardous accumulation volume.

[0058] The early warning system packages and encapsulates the high-stress rockburst early warning signal and the dangerous accumulation volume to generate a downlink data message. The early warning system uses a two-way data communication link to transmit the downlink data message to the frequency converter controller inside the high-pressure water injection equipment. The high-pressure water injection equipment receives the downlink data message and enters a safe standby state. The early warning system continuously monitors the real-time tunnel construction stage output by the hidden Markov model. When it identifies that the real-time tunnel construction stage changes from the post-blasting redistribution stage to the muck removal stage or the support stage, the early warning system uses this non-drilling operation window at the tunnel face to calculate the decompression parameters and generate a hydraulic fracturing decompression control strategy. When the early warning system detects an independent decompression window where the construction stage enters the preparation stage of the next cycle and the advanced drilling operation has not yet started, the early warning system issues an execution command to the high-pressure water injection equipment and triggers the hydraulic fracturing decompression control logic branch to avoid spatial and safety interference between hydraulic fracturing operations and rock drilling by the rock drilling rig.

[0059] The early warning system receives water injection and depressurization parameters sent by the high-pressure water injection equipment. These parameters consist of water injection pressure, water injection flow rate, and water injection duration. The high-pressure water injection equipment executes water injection operations based on the high-stress rockburst early warning signal, inducing crack propagation within the rock mass.

[0060] The microseismic system monitors and collects acoustic signals generated by crack propagation. This system transmits these signals to the early warning system in real time. The early warning system internally sets a fundamental frequency filtering threshold for mechanical noise from the high-pressure water injection equipment and a high-frequency transient trigger threshold. The high-frequency transient trigger threshold is calibrated based on the peak envelope of the mechanical noise from the detector array when the high-pressure water pump is running under no-load, and is set to be greater than the maximum peak amplitude of the pump's mechanical noise, thus filtering out hardware background noise. The early warning system uses the fundamental frequency filtering threshold to perform notch filtering on the acoustic signals to eliminate low-frequency interference from the pump. It also uses the high-frequency transient trigger threshold to extract the target frequency band signal. Finally, the early warning system extracts the real-time monitored pipeline pressure drop slope and cumulative water injection volume from the data fed back by the high-pressure water injection equipment.

[0061] The early warning system internally sets a pressure drop slope threshold and an effective water injection volume threshold. When the pipeline pressure drop slope is greater than or equal to the pressure drop slope threshold and the cumulative water injection volume is greater than or equal to the effective water injection volume threshold, the early warning system marks the target frequency band signal as an effective signal with fracturing characteristics. The early warning system defines the marked target frequency band signal as a high-frequency effective hydraulic fracturing acoustic emission waveform sequence. The early warning system synchronously stores the water injection and depressurization conditions and the high-frequency effective hydraulic fracturing acoustic emission waveform sequence into the storage module for subsequent call and calculation.

[0062] The early warning system reads the high-frequency effective hydraulic fracturing acoustic emission waveform sequence and the updated three-dimensional matrix of elastic wave propagation velocity stored in the storage module. Considering that the introduction of fluid medium during water injection fracturing will cause local wave velocity attenuation in the rock mass, the early warning system obtains the fluid medium wave velocity reduction coefficient through empirical mapping lookup table calculation based on the water injection pressure and the pre-entered rock mass integrity coefficient, and presets it internally. The early warning system extracts the occurrence time range corresponding to the high-frequency effective hydraulic fracturing acoustic emission waveform sequence. Within the occurrence time range, the early warning system uses the fluid medium wave velocity reduction coefficient to perform deceleration correction on the local nodes of the updated three-dimensional matrix of elastic wave propagation velocity, generating a hydraulic fracturing transient wave velocity matrix. The early warning system uses the hydraulic fracturing transient wave velocity matrix to perform three-dimensional spatial positioning calculation on the high-frequency effective hydraulic fracturing acoustic emission waveform sequence, and calculates and generates an acoustic emission spatial coordinate point cloud.

[0063] The early warning system calls the 3D Alpha-Shape algorithm to perform boundary wrapping operations on the acoustic emission spatial coordinate point cloud. The early warning system extracts the boundary nodes inside the acoustic emission spatial coordinate point cloud located on the outer contour. The early warning system connects the boundary nodes to each other and constructs a closed polyhedron structure that fits the concave and convex shape of the actual point cloud.

[0064] The early warning system calculates the spatial volume parameters inside the closed polyhedron structure. The early warning system defines the combination of the closed polyhedron structure and the spatial volume parameters as the actual pressure relief wave volume. The early warning system saves the actual pressure relief wave volume to the storage module for subsequent data interaction and retrieval.

[0065] The early warning system reads the actual depressurization wave volume and the hazardous accumulation volume stored in the storage module. It then calculates the spatial intersection volume by measuring the overlap between the actual depressurization wave volume and the hazardous accumulation volume. The system extracts the original spatial volume corresponding to the hazardous accumulation volume. Finally, it performs a division operation between the spatial intersection volume and the original spatial volume corresponding to the hazardous accumulation volume to obtain the hazardous area coverage rate. The early warning system calculates the hazardous area coverage rate according to the following formula: ; In the formula, Represents the coverage rate of dangerous areas; Represents the volume of the intersection of spaces; This represents the volume of a potentially dangerous accumulation.

[0066] The early warning system internally sets a safety coverage threshold. It compares the coverage rate of the hazardous area with this threshold. When the hazardous area coverage rate is greater than or equal to the threshold, the system generates a pressure relief signal and stores it in the storage module. Simultaneously, the system updates the status flag of the hazardous accumulation volume in the storage module to "safety deactivated" and masks it during subsequent spatial cross-comparison calculations in the operational cycle. This releases system memory resources and prevents the algorithm from triggering duplicate early warning decisions. When the coverage rate of the hazardous area is less than the safe coverage rate threshold, the early warning system generates a secondary pressure relief command and sends it to the high-pressure water injection equipment. The high-pressure water injection equipment receives the command and generates a pipeline connection confirmation prompt at the on-site operation terminal. After receiving an interactive feedback signal confirming that the pipeline is locked to the original water injection hole, the high-pressure water injection equipment triggers the re-injection logic branch. The early warning system continuously monitors the real-time pipeline pressure after the re-injection logic branch is executed and internally presets a pressure holding limit time threshold. When the early warning system determines that the high-pressure water injection equipment cannot maintain the injection pressure or that the coverage rate of the hazardous area after re-injection is still less than the safe coverage rate threshold, the early warning system generates an equipment withdrawal signal and sends it to the high-pressure water injection equipment. After receiving the withdrawal confirmation signal from the high-pressure water injection equipment, the early warning system initiates the hole filling strategy logic branch. The early warning system reads the hazardous accumulation volume and the spatial intersection volume from the storage module, extracts the three-dimensional spatial boundary coordinates corresponding to the hazardous accumulation volume, and extracts the three-dimensional intersection boundary coordinates corresponding to the spatial intersection volume.

[0067] The early warning system obtains the uncovered danger area by subtracting the three-dimensional intersection boundary coordinates from the three-dimensional spatial boundary coordinates within the tunnel's three-dimensional spatial model. The early warning system then extracts the three-dimensional spatial nodes within the uncovered danger area to generate an uncovered danger point cloud.

[0068] The early warning system uses the K-means clustering algorithm to segment the uncovered danger point cloud into regional clusters and calculates the spatial bounding box center of each cluster region to obtain the coordinates of multiple three-dimensional internal target points. The early warning system extracts the current face contour node in the tunnel's three-dimensional spatial model and calculates the shortest spatial line connecting the face contour node to the coordinates of the three-dimensional internal target points. The early warning system defines the intersection of the shortest spatial line and the face contour as the hole opening coordinates and interprets the direction and length of the shortest spatial line as the drilling rig attitude angle and the target drilling depth. The early warning system combines the hole opening coordinates, drilling rig attitude angle, and target drilling depth to generate hole spatial trajectory parameters.

[0069] The early warning system extracts the pre-set kinematic limit pose parameters of the drill arm from the rock drilling rig. It then uses these parameters to perform boundary check calculations on the generated drill rig pose angles. When the drill rig pose angles exceed the pre-set kinematic limit pose parameters, the system increases the number of clusters within the uncovered danger point cloud and re-executes target point addressing and hole filling trajectory calculation until the resolved drill rig pose angles fully meet the kinematic constraints of the rock drilling rig. If the number of clusters reaches a preset upper limit threshold and a drill rig pose angle that meets the kinematic constraints still cannot be obtained, the system determines that the uncovered danger area is a blind spot inaccessible to the rig's machinery. The system then generates a manual hole filling prompt signal and terminates the current hole filling trajectory calculation loop.

[0070] The early warning system has a preset unit volume reference hole layout coefficient. The early warning system extracts the three-dimensional spatial volume parameters corresponding to each cluster region. The early warning system performs a multiplication operation on the three-dimensional spatial volume parameters, the unit volume reference hole layout coefficient and the spatial node distribution density of each cluster. The early warning system performs an integer-rounding operation on the result of the multiplication operation and outputs the rounded value as the number of supplementary boreholes.

[0071] The early warning system combines the spatial trajectory parameters of the supplementary boreholes with the number of supplementary boreholes to generate a supplementary borehole drilling command. The system then sends this command to the drilling rig via a data communication link. The drilling rig receives the command and executes the positioning and supplementary borehole operation. After the positioning and supplementary borehole operation is completed, the early warning system controls the drilling rig to leave the site and issues a supplementary water injection command to the high-pressure water injection equipment for the aforementioned borehole opening coordinates. Simultaneously, the system collects the high-frequency effective hydraulic fracturing acoustic emission waveform sequence generated during the supplementary water injection process, recalculates and accumulates the actual pressure relief wave volume, and repeatedly executes the comparison logic for the dangerous area coverage rate until the dangerous area coverage rate is greater than or equal to the safe coverage rate threshold, thereby ensuring that the pressure relief blind zone is completely eliminated.

[0072] The early warning system receives the real-time tunnel construction phase output by the Hidden Markov Model. When the real-time tunnel construction phase transitions to the drilling phase of the next work cycle, the early warning system activates its spatial coordinate monitoring logic. The system acquires the 3D spatial trajectory coordinates of the drill string for the new round of radial pressure relief hole or exploratory hole operations sent by the drilling rig. It also reads the 3D spatial boundary coordinates corresponding to the hazardous accumulation volume left over from the previous work cycle from the storage module. The system performs a spatial cross-comparison calculation between the 3D spatial trajectory coordinates of the new round of radial pressure relief hole or exploratory hole operations and the 3D spatial boundary coordinates of the previous work cycle. When the 3D spatial trajectory coordinates of the new round of radial pressure relief hole or exploratory hole operations extend through and fall within the spatial range defined by the 3D spatial boundary coordinates, the early warning system generates a spatial re-entry determination signal and stores it in the storage module.

[0073] The early warning system reads the spatial reentry determination signal stored inside the storage module. When the early warning system recognizes the spatial reentry determination signal, it starts the parameter multiplexing acquisition logic.

[0074] The early warning system receives operating machinery parameters transmitted from the drilling rig. From these parameters, it extracts the drilling rig's thrust, rotational speed, rotational torque, and penetration rate. It also extracts the borehole cross-sectional area stored in the storage module. Using the borehole cross-sectional area and operating machinery parameters, the system calculates the actual drilling energy ratio. Next, the system reads the new round of drill string 3D spatial coordinates and correlates these coordinates to generate a posterior spatial energy ratio node. Finally, the system aggregates and arranges these posterior spatial energy ratio nodes in chronological order to generate a posterior energy ratio data sequence, which is then stored in the storage module for later retrieval.

[0075] The early warning system reads the continuous data stream stored in the storage module that records the historical stages of the high-pressure water injection and depressurization operation. The early warning system extracts the drilling specific energy and the three-dimensional spatial coordinates of the drill string from the continuous data stream, thereby providing an original reference benchmark for evaluating the effect of water injection and depressurization. The early warning system defines the drilling specific energy as the initial estimated drilling specific energy.

[0076] The early warning system reads the posterior energy ratio data sequence stored in the storage module. From this sequence, it extracts the new round of drill string 3D spatial coordinates and the actual drilling energy ratio. The system then calls the co-kriging 3D spatial interpolation algorithm to perform spatial interpolation on the historical drill string 3D spatial coordinates and their associated drilling energy ratios recorded in the continuous data stream. This constructs a prospective energy ratio 3D matrix covering the rock mass to be excavated. Using the new round of drill string 3D spatial coordinates, the system performs a mapping addressing operation within the prospective energy ratio 3D matrix, extracting the initial estimated drilling energy ratio associated with these coordinates. Finally, the system performs a subtraction operation between the actual drilling energy ratio and the initial estimated drilling energy ratio to obtain the drilling energy ratio residual. The system calculates this residual using the following formula.

[0077] ; In the formula, Represents the residual drilling specific energy; This represents the actual drilling energy. This represents the initial estimated drilling energy ratio.

[0078] The early warning system internally sets an expected threshold for energy reduction. It compares the residual drilling energy with this threshold. When the residual energy is greater than or equal to the threshold, the system determines that the actual drilling energy has been effectively reduced, indicating that the water injection pressure relief effect in the area is satisfactory, and maintains the current water injection pressure relief parameters of the high-pressure water injection equipment. When the residual energy is less than the threshold, it indicates that the water injection fracturing effect is poor, resulting in intact rock mass and a still high actual drilling energy. The system determines that the actual drilling energy has not been effectively reduced, and that there are invalid fractures within the spatial intersection volume. It then generates a bottom-level linkage correction command and triggers two bottom-level linkage control logics. The system extracts the water injection pressure relief parameters stored in the storage module and calculates the water injection pressure increment using the expected energy reduction threshold and the residual drilling energy. The system calculates the water injection pressure increment according to the following formula.

[0079] ; In the formula, This represents the increase in water injection pressure; Represents the pressure compensation coefficient; This represents the expected threshold for energy reduction; This represents the residual drilling energy.

[0080] The early warning system extracts the maximum safe pressure limit threshold of the high-pressure water injection equipment internally. The maximum safe pressure limit threshold is not a fixed constant, but a hardware protection extreme value dynamically generated by the early warning system after reading the rated maximum head pressure parameter of the hydraulic pump on the factory nameplate of the current high-pressure water injection equipment through the equipment data bus and adjusting the safety redundancy coefficient by 10%-15%.

[0081] The early warning system adds the current water injection pressure within the water injection and depressurization operating parameters to the calculated water injection pressure increment to obtain the expected water injection pressure. The system then compares this expected water injection pressure with the equipment's maximum safe pressure limit threshold. When the expected water injection pressure exceeds the equipment's maximum safe pressure limit threshold, the early warning system triggers the equipment's saturation upper limit explosion-proof protection mechanism. The system halts the accumulation of water injection pressure increments and forcibly assigns the expected water injection pressure to the equipment's maximum safe pressure limit threshold. Simultaneously, the system generates a hole-filling request and jumps to the hole-filling strategy logic branch. The system also determines... The current rock mass at the working face presents an extremely high risk of rockburst. The early warning system immediately issues an emergency speed-reduction and pressure-reduction drilling command or a personnel and equipment withdrawal command to the drilling rig, terminating the current work cycle and generating a manual intervention early warning message. When the expected water injection pressure is less than or equal to the maximum safe pressure limit of the equipment, the early warning system packages the water injection pressure increment and water injection depressurization parameters into a water injection parameter correction data packet. The early warning system uses a two-way data communication link to send the water injection parameter correction data packet to the high-pressure water injection equipment. The high-pressure water injection equipment receives the water injection parameter correction data packet and adjusts the water injection pressure output value.

[0082] The early warning system reads the high-frequency transient trigger threshold stored in the storage module and sets an acoustic sensitivity compensation factor internally. The system then performs a division operation between the high-frequency transient trigger threshold and the acoustic sensitivity compensation factor to obtain a downgraded updated transient trigger threshold. This updated threshold is used to capture extremely weak acoustic emission signals from secondary microcracks within insufficiently depressurized rock masses. The system replaces the high-frequency transient trigger threshold with the updated threshold and stores it in the storage module.

[0083] The early warning system extracts the three-dimensional spatial coordinates of the new drilling tool and its actual drilling energy ratio for conventional excavation areas that have not undergone high-pressure water injection and depressurization operations. The early warning system uses the three-dimensional spatial coordinates of the new drilling tool to perform mapping addressing within the look-ahead energy ratio three-dimensional matrix to extract the corresponding initial estimated drilling energy ratio. The early warning system performs a subtraction operation on the actual drilling energy ratio and the initial estimated drilling energy ratio of the conventional excavation area to obtain the conventional benchmark energy ratio residual. The early warning system uses the residual of conventional benchmark energy to calculate a global spatial variation penalty factor. The system then uses this factor to perform iterative correction on the variogram parameters of the co-kriging 3D spatial interpolation algorithm. The system extracts the residual of conventional benchmark energy from multiple consecutive historical work cycles, including the current time point, to construct a sliding data window. Using the mean square error (MSE) of the data within the sliding data window as the loss function, the system constructs an optimization model. With the goal of minimizing this loss function, the global spatial variation penalty factor is used as the iterative learning step size. The numerical derivative method is employed to calculate the approximate gradient of the loss function with respect to the variogram parameters. A parameter update damping factor is introduced to perform a smoothing filter operation on this approximate gradient. Gradient descent is then used to algebraically update the preset variogram parameters within the co-kriging 3D spatial interpolation algorithm along the negative direction of the smoothed approximate gradient. This completes the closed-loop self-correction of the system's underlying spatial prediction logic.

[0084] To prevent closed-loop oscillation coupling between control parameters and algorithm derivation parameters, the early warning system temporarily freezes the algorithm model update for the water injection affected area during the current operation cycle when sending water injection parameter correction data packets. In subsequent operation cycles where the water injection parameters are determined to be stable, the early warning system uses the drilling energy residual to calculate a local fracture interference penalty factor. The early warning system then calls the local fracture interference penalty factor to perform iterative correction operations on the local interpolation weight coefficients of the prospective energy three-dimensional matrix. The early warning system uses minimizing the absolute value of the drilling energy residual as the objective function, and uses the local fracture interference penalty factor as the iterative learning step size. The early warning system extracts local spatial grid nodes within a preset radius of the new round of drilling tool three-dimensional spatial coordinates to construct a dynamic computational bounding box. Within this dynamic computational bounding box, the system uses the numerical derivative method to calculate the approximate gradient and uses the gradient descent method to perform algebraic updates on the local interpolation weight coefficients of the water injection fracture interference area inside the prospective energy three-dimensional matrix. The early warning system completes the closed-loop self-correction process of the system depressurization assessment logic.

[0085] The intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method is composed of a data acquisition unit, a data processing unit, a data storage unit, a communication and interaction unit, and an engineering execution unit.

[0086] The data acquisition unit interfaces with the drilling parameter acquisition system and the microseismic system. The drilling parameter acquisition system, installed inside the drilling rig, is responsible for extracting operational machinery parameters, while the microseismic system, installed inside the tunnel site, is responsible for capturing microseismic event waveforms and acoustic signals. The data processing unit interfaces with the central processing unit inside the early warning system. The data processing unit receives the underlying monitoring parameters acquired by the data acquisition unit and performs core logic parsing calculations. The data processing unit completes the calculation and control process for drilling energy conversion, real-time tunnel construction stage identification, hazard accumulation volume delineation, and cross-comparison evaluation of pressure relief space.

[0087] The data storage unit interfaces with the storage module within the early warning system. It receives and stores continuous data streams output by the data processing unit, the three-dimensional spatial coordinates of microseismic events, the actual pressure relief wave volume, and the drilling specific energy residual, ready for retrieval. The communication interaction unit establishes a data transmission network and a two-way data communication link. This unit is responsible for maintaining information exchange and flow between the various functionally independent entities, as well as distributing control commands.

[0088] The engineering execution unit connects to the rock drilling rig and high-pressure water injection equipment. The engineering execution unit receives the hole filling drilling instructions and water injection parameter correction data packets transmitted by the communication interaction unit and triggers mechanical operation actions at the tunnel construction site.

[0089] The computer device internally includes a processor, memory, a communication bus, input interfaces, and output interfaces. The communication bus establishes a data flow channel within the computer device, connecting the processor, memory, input interfaces, and output interfaces. The memory is divided into a program execution area and a data storage area. The memory loads control instruction programs in the program execution area, and stores computational variables and output parameters generated during the early warning system's operation in the data storage area. The processor calls the communication bus to read the control instruction programs loaded in memory, executes the control instruction programs, and triggers the underlying computational and judgment logic of the early warning system.

[0090] The input interface connects to the field monitoring equipment, receiving the acquired parameters transmitted by the equipment and importing them into the processor for parsing. The output interface connects to the engineering execution unit, receiving control commands generated by the processor and transmitting them to the engineering execution unit to trigger mechanical actions. A computer-readable storage medium contains instruction set code; the processor calls and executes this code to trigger the computational control steps of the intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method.

[0091] Computer-readable storage media are divided into non-volatile media and volatile media. Non-volatile media are constructed using read-only memory, solid-state drives, or magnetic hard disk drives, while volatile media are constructed using random access memory. Computer-readable storage media interface with the internal communication bus of a computer device, establishing a data exchange channel with the processor via the communication bus. The processor extracts and reads instruction set code from the computer-readable storage media, executes the instruction set code, and completes the low-level logic deduction and control information distribution process.

[0092] Specific application examples: To further aid in understanding the specific implementation logic and working principle of the present invention, the following provides specific application examples and experimental verification results based on the daily tunneling scenario of a deep-buried tunnel under high ground stress in Southwest China.

[0093] In this application scenario, the intelligent early warning and prevention system for high-stress rockburst using the drilling and blasting method of this invention is needed to assess the rockburst risk in the rock mass ahead of the tunnel face, and to perform pressure relief and explosion prevention operations using high-pressure water injection equipment, while simultaneously verifying the system's underlying parameter self-correction and linkage control capabilities. To illustrate the data calculation and state assessment process within the system, the following explanation substitutes the entity test parameters from the scenario into the calculation module within the early warning system.

[0094] When the Hidden Markov Model outputs the drilling stage as the real-time stage of tunnel construction, the early warning system initiates the logic for extracting operating machinery parameters. The early warning system extracts the operating machinery parameters transmitted by the drilling rig and preset parameters, including the known drilling rig thrust. The drilling rig rotation speed is 15000N. The drilling rig's rotational torque is 3 r / s. The drilling rig penetration rate is 500 N·m. The speed is 0.02 m / s, and the borehole cross-sectional area is... It is 0.005m 2 Pi constant Take 3.14. The early warning system calculates and generates the drilling specific energy. : ; The drilling energy will be input into the rock mechanics mapping table to update the three-dimensional matrix of elastic wave propagation velocity.

[0095] As the tunnel construction transitions from the real-time phase to the post-blasting redistribution phase, the early warning system utilizes microseismic event waveform signals acquired by the microseismic system for three-dimensional spatial positioning and delineates the hazardous accumulation volume. Subsequently, the system instructs high-pressure water injection equipment to perform hydraulic fracturing and depressurization. After water injection is completed, the system calculates the actual depressurization swept volume and extracts the spatial intersection volume generated by the overlap between the actual depressurization swept volume and the hazardous accumulation volume. 120m 3 Known hazardous accumulation volume 150m 3 The early warning system calculates the coverage rate of dangerous areas. : ; Assuming the internally set security coverage threshold is 0.85, due to If (0.80) is less than the safety coverage threshold (0.85), the early warning system generates a secondary depressurization command or initiates the hole filling strategy logic branch.

[0096] Upon entering the drilling phase of the next work cycle, the early warning system initiates spatial re-entry detection and collects parameter reuse logic. The system then extracts data from the drilling rig again to calculate the actual drilling energy. The value is 72.5 MPa. The system extracts the corresponding initial estimated drilling specific energy from the three-dimensional look-ahead specific energy matrix. (That is, the baseline value before depressurization in the previous cycle, which is 97.2 MPa in this case). The early warning system calculates the drilling specific energy residual. : ; The warning system has an internally set threshold for the expected decrease in specific energy. It is 35.0 MPa. Because... Less than (35.0 MPa), the early warning system determines that there are invalid cracks within the spatial intersection volume. The set pressure compensation coefficient is known. The value is 0.4. The early warning system calculates the injection pressure increment for the next compensation. : ; The early warning system packages the water injection pressure increment and water injection depressurization parameters into a water injection parameter correction data package, which is then sent to the high-pressure water injection equipment to adjust the subsequent water injection pressure output.

[0097] To demonstrate the effectiveness of the intelligent early warning and control system for high-stress rockburst using the drill-and-blast method of the present invention, data from 10 consecutive tunneling cycles were collected at the engineering site for a comparative experiment. The experiment compared the actual drilling energy change trend at the working face after using the traditional fixed-parameter water injection depressurization method and the intelligent early warning and control system (closed-loop self-correction) of the present invention.

[0098] The conclusions are as follows: Reference Appendix Figure 2 , attached Figure 2 The solid line with a square mark represents the actual drilling energy ratio using the traditional fixed-parameter water injection and depressurization method, while the solid line with a circular mark represents the actual drilling energy ratio using the intelligent early warning and prevention system of this invention.

[0099] From the appendix Figure 2 It is evident that traditional fixed-parameter methods, unable to detect the stress relief blind zone within the rock mass, consistently maintain the actual drilling specific energy at a high level above 90 MPa, indicating that the risk of high-stress rockburst remains unresolved. However, by employing the intelligent early warning and control system of this invention, the system continuously monitors microseismic event waveform signals and calculates the coverage rate of the dangerous area. It then uses the drilling specific energy residual to trigger the bottom-layer linkage control logic (outputting an incremental water injection pressure). This allows the actual drilling specific energy to rapidly and stably decrease underground to a safe range of approximately 60 MPa with increasing cycle count. This demonstrates that the system of this invention, through multi-source data fusion and closed-loop self-correction, can eliminate stress relief blind zones and reduce the integrity and high stress concentration of the rock mass ahead of the tunnel face.

Claims

1. A smart early warning and prevention system for high-stress rockburst using the drill-and-blast method, characterized in that, This includes an early warning system, a rock drilling rig, a micro-vibration system, and high-pressure water injection equipment; The rock drilling rig is used to acquire the operating machinery parameters and the three-dimensional spatial coordinates of the drill bit, and to send the operating machinery parameters and the three-dimensional spatial coordinates of the drill bit to the early warning system; The microseismic system is used to collect tunnel environmental noise data and microseismic event waveform signals, and transmit the tunnel environmental noise data and the microseismic event waveform signals to the early warning system in real time. The early warning system is used to receive the parameters of the operating machinery, the three-dimensional spatial coordinates of the drilling tool, the tunnel environmental noise data, the waveform signal of the micro-vibration event, the pipeline pressure drop slope and cumulative water injection volume fed back by the high-pressure water injection equipment, perform multi-source data fusion calculation and three-dimensional spatial matching calculation, and output control commands to the high-pressure water injection equipment to form a data communication closed loop link. The high-pressure water injection equipment is used to monitor the pipeline pressure drop slope and the cumulative water injection volume in real time, and feeds back the pipeline pressure drop slope and the cumulative water injection volume to the early warning system. It receives the control command output by the early warning system and performs water injection operation in the danger zone corresponding to the three-dimensional spatial coordinates of the drill bit according to the high ground stress rockburst early warning signal generated by the early warning system.

2. The intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method according to claim 1, characterized in that, The early warning system is equipped with a calculation module and a storage module. The storage module pre-stores a hidden Markov model, a rock mechanics mapping table, and a three-dimensional matrix of elastic wave propagation velocity. The early warning system inputs the operating machinery parameters and the tunnel environmental noise data into the hidden Markov model, and the hidden Markov model outputs the real-time stage of tunnel construction.

3. The intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method according to claim 2, characterized in that, When the tunnel construction is in the drilling stage, the early warning system extracts the parameters of the operating machinery and calculates and generates the drilling specific energy. The early warning system then inputs the drilling specific energy into the rock mechanics mapping table. The rock mechanics mapping table outputs a uniaxial compressive strength distribution sequence, and the early warning system uses the uniaxial compressive strength distribution sequence to update the three-dimensional matrix of elastic wave propagation velocity, generating an updated three-dimensional matrix of elastic wave propagation velocity.

4. The intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method according to claim 3, characterized in that, When the tunnel construction real-time stage is in the post-blasting redistribution stage, the early warning system calls the updated elastic wave propagation velocity three-dimensional matrix to perform three-dimensional spatial positioning calculation on the microseismic event waveform signal to obtain the three-dimensional spatial coordinates of the microseismic event. The early warning system performs integral calculations on the waveform signal of the microseismic event to extract the energy released by the microseismic event, and calculates the spatial energy density of the microseismic event by combining the three-dimensional spatial coordinates of the microseismic event. When the energy density of the microseismic space is greater than the preset energy density critical threshold, the early warning system generates a high-stress rockburst early warning signal. The early warning system establishes a three-dimensional spatial model of the tunnel and calls a spatial density clustering algorithm to perform spatial point cloud aggregation operation on the three-dimensional spatial coordinates of the microseismic events that exceed the energy density critical threshold. The system calculates, generates, and marks the microseismic event cluster area within the three-dimensional spatial model of the tunnel, and defines the microseismic event cluster area as a dangerous cluster volume.

5. The intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method according to claim 4, characterized in that, The early warning system monitors the pipeline pressure drop slope and the cumulative water injection volume fed back by the high-pressure water injection equipment after receiving the high-stress rockburst early warning signal and performing water injection operation, and extracts the high-frequency effective hydraulic fracturing acoustic emission event generated by the water injection operation from the micro-vibration event waveform signal. The early warning system calculates the envelope three-dimensional space formed by the high-frequency effective hydraulic fracturing acoustic emission event and defines the envelope three-dimensional space as the actual depressurization wave volume.

6. The intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method according to claim 5, characterized in that, The early warning system uses the dangerous accumulation volume and the actual depressurization wave volume to calculate and output the dangerous area coverage rate; The early warning system issues the control command to the high-pressure water injection equipment based on the coverage of the dangerous area, and extracts and generates the intersection area based on the dangerous accumulation volume and the actual depressurization wave volume.

7. The intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method according to claim 6, characterized in that, The early warning system continuously monitors the real-time stage changes of the tunnel construction. When the three-dimensional spatial coordinates of the drilling tool fall into the intersection area, the early warning system re-collects the operating machinery parameters and generates the actual drilling energy ratio. The early warning system extracts the initial estimated drilling ratio energy stored in the storage module, which is generated by spatial interpolation of the drilling ratio energy, and calculates the drilling ratio energy residual between the actual drilling ratio energy and the initial estimated drilling ratio energy. The early warning system triggers a bottom-level linkage control operation based on the drilling energy residual. The early warning system then reverses the rock mechanics mapping table and updates the updated elastic wave propagation velocity three-dimensional matrix to complete the bottom-level logic closed-loop self-correction.

8. The intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method according to claim 7, characterized in that, When the early warning system triggers the underlying linkage control operation based on the drilling specific energy residual, it is specifically used for: The early warning system extracts the water injection and depressurization operating parameters stored in the storage module, and calculates the water injection pressure increment using the internally preset specific energy decrease expected threshold and the drilling specific energy residual. The early warning system adds the water injection pressure included in the water injection and depressurization working condition parameters to the calculated water injection pressure increment to obtain the expected water injection pressure. The early warning system then compares the expected water injection pressure with the internally preset maximum safe pressure limit threshold of the equipment.

9. The intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method according to claim 8, characterized in that, The early warning system compares the expected water injection pressure with the maximum safe pressure limit threshold of the equipment, and is specifically used for: When the expected water injection pressure exceeds the maximum safe pressure limit threshold of the equipment, the early warning system triggers the equipment saturation upper limit explosion protection mechanism. The early warning system cuts off the accumulation of the water injection pressure increment and assigns the expected water injection pressure to the maximum safe pressure limit threshold of the equipment. When the expected water injection pressure is less than or equal to the maximum safe pressure limit threshold of the equipment, the early warning system packages the water injection pressure increment and the water injection depressurization condition parameters to generate a water injection parameter correction data packet. The early warning system uses the data communication closed-loop link to send the water injection parameter correction data packet to the high-pressure water injection equipment. The high-pressure water injection equipment receives the water injection parameter correction data packet and adjusts the water injection pressure output value.

10. The intelligent early warning and prevention system for high-stress rockburst using the drill-and-blast method according to claim 9, characterized in that, The early warning system is also used for: The high-frequency transient trigger threshold stored in the storage module is read, and the early warning system internally sets an acoustic sensitivity compensation factor. The early warning system performs a division operation between the high-frequency transient trigger threshold and the acoustic sensitivity compensation factor to obtain a downgraded updated transient trigger threshold. The early warning system replaces the high-frequency transient trigger threshold with the updated transient trigger threshold and stores the updated transient trigger threshold in the storage module.

Images