Real-time monitoring system for surrounding rock of semi-coal rock roadway based on optical fiber sensing technology

Abstract

The application relates to the technical field of coal mine safety monitoring and disaster early warning, and discloses a real-time monitoring system for surrounding rock of a semi-coal rock roadway based on optical fiber sensing technology, which comprises the following steps: a data acquisition module acquires original strain data of a borehole optical fiber; a spatial signal reconstruction module performs spatial smoothing fitting to output reconstructed data; a lithology interface dynamic calibration module identifies a coal-rock boundary depth and calculates an initial stiffness ratio; a spatial shear characteristic extraction module calculates a shear decoupling index and triggers a signal when the index exceeds a limit; a time-domain rheological characteristic extraction module calculates a rheological decoupling degree in response to the trigger; and a multi-dimensional coupling criterion module fuses the initial stiffness ratio, the shear decoupling index and the rheological decoupling degree to calculate a comprehensive risk index and output a result. The spatial signal reconstruction module performs spatial domain smoothing fitting processing on the original strain data, can effectively extract local spatial geometric distortion characteristics of the surrounding rock, and acutely captures micro-bending deformation of a coal-rock interface in an initial stage of dislocation while smoothing away data fluctuation interference.

Description

Technical Field

[0001] This invention relates to the field of coal mine safety monitoring and disaster early warning technology, specifically a real-time monitoring system for the surrounding rock of semi-coal and rock roadways based on fiber optic sensing technology. Background Technology

[0002] In the mining of deep coal resources, semi-coal-rock roadways are a highly unique and common engineering structure. Their roofs consist of both soft coal seams and hard rock strata. This complex structure, with its alternating layers of soft and hard rock, results in extremely weak interlayer bonding, making them highly susceptible to interlayer slippage and shear slippage under geostress, which can lead to roof collapse accidents. To ensure operational safety underground and prevent casualties and equipment damage caused by surrounding rock disasters, it is essential to monitor these hidden interlayer structural changes in real time and with precision. This allows for the detection of abnormal signals in the early stages of a potential disaster, providing valuable time for reinforcement or evacuation of engineers.

[0003] Existing rock mass monitoring technologies primarily rely on point sensors and conventional distributed monitoring methods. Multi-point displacement gauges, delamination gauges, and borehole stress gauges are widely installed in boreholes at key sections of roadways to acquire data on rock displacement or stress changes at specific depths. With advancements in fiber optic sensing technology, distributed fiber optic systems are also being applied to mine monitoring. By embedding sensing fibers into boreholes or surfaces, the principle of optical time-domain reflectometry is used to acquire strain information distributed along the path. Compared to traditional point sensors, this technology significantly increases the density of monitoring data, enabling continuous recording of the overall convergent deformation trend of the surrounding rock in roadways, and playing a positive role in assessing the macroscopic stability of the surrounding rock. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a real-time monitoring system for the surrounding rock of semi-coal and rock roadways based on fiber optic sensing technology. This system solves the problems in existing technologies, such as the tendency of general filtering algorithms to filter out minute local shear precursor signals as noise, the inability of a single threshold criterion to distinguish between transient mechanical disturbances and substantial rheological damage to the surrounding rock, and the lack of geological stiffness parameter correction, which leads to poor early warning targeting and high false alarm rate in complex environments with alternating soft and hard layers.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a real-time monitoring system for surrounding rock in semi-coal and rock roadways based on fiber optic sensing technology, comprising: The data acquisition module is used to acquire raw strain data from the distributed fiber optic sensing device inside the surrounding rock borehole from the fiber optic demodulation device. The spatial signal reconstruction module is used to perform spatial domain smoothing fitting on the original strain data to output reconstructed strain data. The lithological interface dynamic calibration module is used to identify the coal-rock boundary depth based on the spatial gradient distribution of the reconstructed strain data and to calculate the initial stiffness ratio. The spatial shear feature extraction module is used to calculate the spatial strain curvature at the coal-rock boundary depth based on the reconstructed strain data, and to calculate the shear decoupling index based on the spatial strain curvature. When the shear decoupling index exceeds a spatial threshold, a trigger signal is generated. The time-domain rheological feature extraction module is used to respond to the trigger signal and calculate the strain aging factor on both sides of the coal-rock boundary depth, and calculate the rheological decoupling degree based on the strain aging factor on both sides. The multidimensional coupling criterion module is used to calculate a comprehensive risk index based on the initial stiffness ratio, the shear decoupling index, and the rheological decoupling degree, and to output a risk warning command based on the comprehensive risk index.

[0006] Preferably, the data acquisition module is specifically used for: Establish a spatial coordinate axis with the borehole opening as the origin and the direction of borehole depth as the positive direction, and a physical time coordinate axis with the monitoring start time as the starting point; The spatial location of the reflection point is located using optical time-domain reflectometry, and the axial strain value at each point along the optical fiber is calculated by using the linear mapping relationship between the frequency shift characteristic of the backscattered light signal and the axial strain. The calculated axial strain value is mapped to a spatiotemporal coordinate system composed of the spatial coordinate axis and the physical time coordinate axis to generate the original strain data.

[0007] Preferably, the spatial signal reconstruction module is specifically used for: For each sampling time, the physical length of the sliding window is set to cover the expected inter-layer shear influence domain of coal and rock, and the half-width of the sliding window is calculated based on the spatial sampling interval. A sliding window convolution operation is performed on the original strain data along the spatial coordinate axis. For the sampling point located at the center point of the sliding window, the original data points before and after the center point are selected and weighted and summed to obtain the single-point reconstructed strain value corresponding to the center point. The reconstructed strain data is composed of the single-point reconstructed strain values ​​corresponding to all sampling positions and is then output.

[0008] Preferably, the dynamic calibration module for lithological interfaces is specifically used for: The spatial strain gradient along the borehole depth direction of the reconstructed strain data is calculated using the central difference algorithm. The spatial strain gradient is scanned along its entire length to identify local maxima. The peak position with the smallest distance is selected as the final coal-rock boundary depth in combination with a preset reference boundary depth. Based on the coal-rock boundary depth, the monitoring area is spatially divided into a first spatial subdomain and a second spatial subdomain, and the average strain value in each spatial subdomain is calculated. The ratio of the average strain value in the first spatial subdomain to that in the second spatial subdomain is defined as the initial stiffness ratio, which is a dimensionless number.

[0009] Preferably, the spatial shearing feature extraction module is specifically used for: The second-order central difference algorithm is used to calculate the coal-rock boundary depth and the spatial strain curvature within the neighborhood of the coal-rock boundary depth. A rock mass domain was selected as a geological reference area, and the standard deviation of the background noise of the curvature distribution within the rock mass domain was calculated. Calculate the ratio of the peak curvature at the coal-rock boundary depth to the standard deviation of the background noise, and define the ratio as the shear decoupling index; The shear decoupling index, calculated in real time, is compared with a spatial threshold. If the shear decoupling index is greater than the spatial threshold, it is determined that spatial shearing signs have appeared at the coal-rock interface.

[0010] Preferably, the time-domain rheological feature extraction module is specifically used for: Define a rheological time window as the time step for sparsification, and extract the reconstructed strain data of the current moment, the historical moment before one rheological time window, and the historical moment before two rheological time windows; Using the second-order backward difference algorithm, the strain acceleration distribution within the entire monitoring range is calculated based on the rheological time window, and the strain acceleration distribution is defined as the strain aging factor. Based on the coal body domain and rock body domain divided by the coal-rock boundary depth, the average strain aging factor in the coal body domain and the average strain aging factor in the rock body domain are calculated respectively. Calculate the absolute value of the difference between the average strain aging factor in the coal body domain and the average strain aging factor in the rock body domain, and define the absolute value of the difference as the rheological decoupling degree.

[0011] Preferably, the multidimensional coupling criterion module is specifically used for: A critical rheological threshold is introduced to standardize the rheological decoupling degree, and standardized rheological indices are calculated. The initial stiffness ratio is adjusted by introducing a geological sensitivity index, and the nonlinear amplification factor is calculated. A weighted Euclidean distance model is constructed, and the weighted sum of squares is calculated for the ratio of the shear decoupling index to the spatial threshold and the ratio of the rheological decoupling degree to the critical rheological threshold, respectively. The weighted sum of squares is square-rooted, and the result is multiplied by the nonlinear amplification factor to obtain the comprehensive risk index.

[0012] Preferably, after the lithological interface dynamic calibration module divides the monitoring area into a first spatial subdomain and a second spatial subdomain based on the coal-rock boundary depth, it is further used for: Based on the physical property that the elastic modulus of coal is less than that of rock mass, the average strain values ​​of the first spatial subdomain and the second spatial subdomain are compared. The spatial subdomain on the side with the larger average strain value is identified as the coal mass domain, and the spatial subdomain on the side with the smaller average strain value is identified as the rock mass domain.

[0013] Preferably, in the weighted Euclidean distance model, spatial feature weights and temporal feature weights are set, and the sum of the spatial feature weights and the temporal feature weights is 1, wherein the spatial feature weights are greater than the temporal feature weights.

[0014] Preferably, when the multidimensional coupling criterion module outputs a risk warning command, it specifically executes the following logic: The calculated comprehensive risk index is compared with a preset risk classification threshold to determine the risk status; Based on the determined risk status, the monitoring parameters are automatically adjusted and / or corresponding early warning signals are output.

[0015] The real-time monitoring system for surrounding rock in semi-coal and rock roadways based on fiber optic sensing technology provided by this invention has the following beneficial effects: This invention uses a spatial signal reconstruction module to perform spatial domain smoothing and fitting on the original strain data, and coordinates with a spatial shear feature extraction module to calculate the spatial strain curvature and shear decoupling index. This effectively extracts the local spatial geometric distortion features of the surrounding rock. Based on the monitoring method of reconstructed strain data and curvature distribution, it can smooth out data fluctuation interference and keenly capture the micro-bending deformation in the early stage of coal-rock interface displacement. Compared with the existing technology that relies on the strain value or displacement of a single measuring point, this invention solves the shortcomings of traditional methods, which make it difficult to detect early local distortions due to the signal being submerged by background noise and are prone to missing small geological anomalies.

[0016] This invention constructs a multi-dimensional coupling criterion that includes spatial morphology and temporal rheological trends. It is only judged as a dangerous situation when the shear decoupling index exceeds the limit and the rheological decoupling degree shows accelerated separation. This effectively distinguishes between substantial geological damage and transient mechanical disturbances. Compared with the existing technology that triggers an alarm by setting a single spatial threshold, this invention solves the problem of misjudging the sudden strain caused by blasting vibration or equipment thermal noise as rock instability, and significantly reduces the false alarm rate of the system.

[0017] This invention introduces a dynamic calibration mechanism for lithological interfaces and calculates the initial stiffness ratio, which is then incorporated as a geological sensitivity coefficient into the weighted calculation of the comprehensive risk index. The warning weight is automatically adjusted based on the actual relative softness and hardness of the coal and rock masses, applying stricter risk assessment standards to areas with weak coal seams. Compared to the extensive management method of using a fixed and uniform alarm threshold for the entire roadway in the existing technology, this invention solves the defects of the monitoring system, which cannot adapt to different geological structural sections and has poor warning targeting and low accuracy in complex environments with alternating soft and hard layers. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the system structure according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the method flow according to an embodiment of the present invention; Figure 3 This is a diagram showing the strain distribution and spatial strain curvature characteristic curves along the borehole depth direction according to an embodiment of the present invention. Figure 4 This is a graph showing the evolution trend of shear decoupling index, rheological decoupling degree, and comprehensive risk index over monitoring time in an embodiment of the present invention.

[0019] Explanation of icon numbers: 10. Distributed optical fiber sensing device; 20. Optical fiber demodulation device; 30. Data processing terminal; 301. Data acquisition module; 302. Spatial signal reconstruction module; 303. Lithological interface dynamic calibration module; 304. Spatial shear feature extraction module; 305. Temporal rheological feature extraction module; 306. Multidimensional coupling criterion module. Detailed Implementation

[0020] 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.

[0021] See Figure 1 , Figure 1 This is a schematic diagram of a real-time monitoring system for the surrounding rock of a semi-coal-rock roadway based on fiber optic sensing technology according to an embodiment of the present invention. The real-time monitoring system for the surrounding rock of a semi-coal-rock roadway based on fiber optic sensing technology provided in this embodiment of the invention mainly includes: a distributed fiber optic sensing device 10, a fiber optic demodulation device 20, and a data processing terminal 30.

[0022] The distributed optical fiber sensing device 10 is deployed inside a borehole in the surrounding rock of a semi-coal-rock roadway. The distributed optical fiber sensing device 10 is fixed within the borehole using a full-length grouting process, where the annular space between the sensing optical fiber and the borehole wall is filled and solidified along its entire length using coupling materials such as cement mortar or resin. This full-length grouting process creates a deformation-coordinated system between the distributed optical fiber sensing device 10 and the surrounding rock. This deformation-coordinated system means that the optical fiber, grouting material, and surrounding rock maintain continuous displacement during stress deformation, allowing the optical fiber to synchronously and without relative slippage sense the true minute strain of the surrounding rock. This is the physical basis for subsequent precise signal processing.

[0023] The fiber optic demodulation device 20 is physically connected to the distributed fiber optic sensing device 10. The fiber optic demodulation device 20 is configured to transmit probe pulses to the distributed fiber optic sensing device 10 and receive backscattered light signals returned by the distributed fiber optic sensing device 10. The fiber optic demodulation device 20 demodulates the backscattered light signals to acquire raw strain data distributed along the length of the distributed fiber optic sensing device 10.

[0024] The data processing terminal 30 is communicatively connected to the fiber optic demodulation device 20. The data processing terminal 30 receives the raw strain data output by the fiber optic demodulation device 20 and performs subsequent data cleaning, feature extraction, and logical operations. Specifically, the data processing terminal 30 may include: a data acquisition module 301, a spatial signal reconstruction module 302, a lithological interface dynamic calibration module 303, a spatial shear feature extraction module 304, a time-domain rheological feature extraction module 305, and a multidimensional coupling criterion module 306.

[0025] The data acquisition module 301 is configured to acquire the current raw strain data from the fiber optic demodulation device 20 at preset time intervals. The raw strain data includes positional information along the borehole depth and the corresponding axial strain value.

[0026] The spatial signal reconstruction module 302 is connected to the data acquisition module 301. The spatial signal reconstruction module 302 is configured to receive the raw strain data and perform spatial domain smoothing fitting on the raw strain data using the Savitzky-Golay filtering algorithm. The advantage of using the Savitzky-Golay filtering algorithm in this embodiment is that, based on the principle of local polynomial least squares fitting, it can effectively filter out high-frequency random white noise while preserving the geometric features of the signal waveform, such as width, height, and inflection points, to the maximum extent possible, preventing the loss of local minute abrupt changes reflecting interlayer faulting due to excessive smoothing. The spatial signal reconstruction module 302 outputs reconstructed strain data that retains local abrupt change features and removes random white noise.

[0027] The lithological interface dynamic calibration module 303 is connected to the spatial signal reconstruction module 302. During the initialization phase of the monitoring system, the lithological interface dynamic calibration module 303 calculates the spatial gradient distribution of the reconstructed strain data. The lithological interface dynamic calibration module 303 determines the boundary depth between the coal seam and the rock strata by identifying the local maxima of the spatial gradient distribution. The boundary depth refers to the specific depth coordinates along the borehole axis where the geological properties physically transform from coal to rock. The lithological interface dynamic calibration module 303 is also configured to calculate the initial stiffness ratio characterizing the geological conditions based on the average strain values ​​on both sides of the boundary depth.

[0028] The spatial shear feature extraction module 304 is connected to the spatial signal reconstruction module 302 and the lithological interface dynamic calibration module 303. During the real-time monitoring phase, the spatial shear feature extraction module 304 calculates the spatial strain curvature near the boundary depth based on the reconstructed strain data. The spatial shear feature extraction module 304 further calculates the shear decoupling index; when the shear decoupling index exceeds a preset spatial threshold, the spatial shear feature extraction module 304 generates a trigger signal.

[0029] The temporal rheological feature extraction module 305 responds to the trigger signal generated by the spatial shear feature extraction module 304. The temporal rheological feature extraction module 305 is configured to retrieve reconstructed strain data from stored historical moments. Based on a preset rheological time window, the temporal rheological feature extraction module 305 calculates the strain aging factors of the coal body and rock mass on both sides of the boundary depth over a long period. By comparing the strain aging factors on both sides, the temporal rheological feature extraction module 305 calculates the rheological decoupling degree.

[0030] The multidimensional coupling criterion module 306 is connected to both the spatial shear feature extraction module 304 and the temporal rheological feature extraction module 305. The multidimensional coupling criterion module 306 receives the shear decoupling index and the rheological decoupling degree. The multidimensional coupling criterion module 306 is configured to apply a multidimensional coupling risk index model for judgment. When both the shear decoupling index and the rheological decoupling degree simultaneously meet preset conditions, the multidimensional coupling criterion module 306 determines that an interlayer shear instability event has occurred. An interlayer shear instability event refers to a dynamic disaster phenomenon involving relative shear slippage, displacement, or structural failure at the weakly bonded portion of the coal-rock interface under the action of surrounding rock stress. The multidimensional coupling criterion module 306 is also configured to calculate a comprehensive risk coefficient and output monitoring results based on the initial stiffness ratio.

[0031] See Figure 2 , Figure 2 This is a flowchart illustrating the overall process of a monitoring method according to an embodiment of the present invention. This embodiment of the present invention provides a real-time monitoring method for the surrounding rock of semi-coal-rock roadways based on fiber optic sensing technology, comprising the following steps: Step S100: Obtain raw strain data from the distributed fiber optic sensing device inside the surrounding rock borehole from the fiber optic demodulation device; Step S200: Perform spatial domain smoothing fitting on the original strain data to output reconstructed strain data; Step S300: Identify the coal-rock boundary depth based on the spatial gradient distribution of the reconstructed strain data, and calculate the initial stiffness ratio; Step S400: Calculate the spatial strain curvature at the coal-rock boundary depth based on the reconstructed strain data, and calculate the shear decoupling index based on the spatial strain curvature. Generate a trigger signal when the shear decoupling index exceeds the spatial threshold. Step S500: Respond to the trigger signal and calculate the strain aging factor on both sides of the coal-rock boundary depth, and calculate the rheological decoupling degree based on the strain aging factor on both sides; Step S600: Calculate the comprehensive risk index based on the initial stiffness ratio, shear decoupling index and rheological decoupling degree, and output a risk warning command based on the comprehensive risk index.

[0032] Specifically, in step S100, the raw strain data distributed along the borehole depth direction by the distributed optical fiber sensing device 10 is collected. The raw strain data includes borehole depth location information distributed along the optical fiber and the corresponding axial strain value. Specifically, in step S200, the Savitzky-Golay filtering algorithm is used to perform polynomial fitting on the original strain data to generate reconstructed strain data that is denoised and retains local features. Specifically, in step S300, the spatial gradient of the reconstructed strain data is calculated, the depth position corresponding to the maximum gradient value is identified to determine the coal-rock boundary depth, and the initial stiffness ratio on both sides of the coal-rock boundary depth is calculated. Specifically, in step S400, the spatial strain curvature and shear decoupling index at the coal-rock boundary depth are calculated based on the reconstructed strain data, and it is determined whether the shear decoupling index exceeds a preset spatial threshold. Specifically, in step S500, when the shear decoupling index exceeds the preset spatial threshold, the strain aging factor on both sides of the coal-rock boundary depth is calculated based on the preset rheological time window, and the difference between the strain aging factors on both sides is calculated to obtain the rheological decoupling degree. If the shear decoupling index does not exceed the preset spatial threshold, it is determined that the current state is safe, and the process returns to step S100 to continue monitoring for the next cycle. Specifically, in step S600, a multidimensional coupling risk index model is applied to jointly determine the shear decoupling index and the rheological decoupling degree. When both meet the preset conditions, an interlayer shear instability warning is output, and the comprehensive risk coefficient is calculated in combination with the initial stiffness ratio.

[0033] Step S100 is the raw strain data acquisition stage, which is mainly performed by the data acquisition module 301. To ensure that the monitoring data can accurately reflect the true stress state of the surrounding rock, in this embodiment, the distributed optical fiber sensing device 10 is pre-laid inside the surrounding rock borehole of the semi-coal-rock roadway and consolidated with the surrounding rock mass through a full-length grouting process. This full-length cemented installation method ensures that the deformation of the surrounding rock can be transmitted to the sensing optical fiber without slippage through the grouting material, so that the strain measured by the optical fiber can directly characterize the actual geological deformation of the surrounding rock.

[0034] During the execution of step S100, the data acquisition module 301 sends an acquisition command to the fiber optic demodulation device 20 via the communication interface. The fiber optic demodulation device 20 transmits probe pulse light to the distributed fiber optic sensing device 10 and receives the backscattered light signal returning along the fiber length. Specifically, the fiber optic demodulation device 20 uses optical time domain reflection (OTDR) to locate the spatial position of the reflection point and uses the linear mapping relationship between Brillouin scattering frequency shift or Rayleigh scattering spectral frequency shift and axial strain to calculate the axial strain value at each point along the fiber.

[0035] The data acquisition module 301 receives the signal output by the fiber optic demodulation device 20 and converts it into a digital signal sequence. To facilitate subsequent data processing and feature extraction, this embodiment establishes a spatial coordinate axis with the borehole opening as the origin and the direction of borehole depth as the positive direction. And the physical time axis starting from the monitoring start time. The data acquisition module 301 maps the parsed data to the spatiotemporal coordinate system to generate the original strain data.

[0036] In this embodiment, the original strain data is represented as a function of spatial position. and time changing function In a discretized digital system, the original strain data is stored as a two-dimensional matrix. For the first... The spatial sampling point and the first The formula for calculating the original strain data at each sampling time point is as follows: ; in, Indicates time At that moment, at the borehole depth The original axial strain value at the location is expressed in microstrain (µs). ); Indicates the first The depth position of each spatial sampling point satisfies ,in The spatial sampling interval is determined by the spatial resolution of the fiber optic demodulation device 20. In this embodiment, The preferred value range is 0.05 meters to 0.5 meters to ensure that the local deformation characteristics of the coal-rock interface can be captured; Indicates the first Each sampling time satisfies ,in The preset time sampling frequency takes into account the gradual changes in the rheological process of the surrounding rock and the suddenness of shear failure. The preferred value range is 0.1Hz to 10Hz; This represents the frequency shift characteristic of the backscattered light signal received by the fiber optic demodulation device 20 at the corresponding spatiotemporal location, specifically the Brillouin frequency shift or Rayleigh scattering frequency shift. This represents the optical-to-mechanical conversion coefficient function, which is a linear coefficient determined by the properties of the optical fiber material, with a typical value of 0.048 MHz / (For Brillouin scattering) or 0.15 GHz / (For Rayleigh scattering), the specific values ​​are calibrated according to the type of fiber selected.

[0037] It should be noted that due to electromagnetic interference in the downhole environment and the thermal noise of the photoelectric equipment itself, the output at this time... This typically includes random white noise of a certain amplitude. While this noise has a relatively small impact in conventional single-point monitoring, it is significantly amplified during subsequent high-order differential operations (such as calculating curvature and acceleration), thus masking the true deformation characteristics. Therefore, it is necessary to transmit the raw strain data to the space signal reconstruction module 302 for subsequent smoothing processing.

[0038] Step S200 is the spatial signal reconstruction stage, which is mainly performed by the spatial signal reconstruction module 302. This is because the raw strain data acquired in step S100... Due to limitations in photoelectric detection principles, the data typically contains random white noise. While this random white noise has limited impact when directly reading strain values, its nature amplifies the noise significantly during subsequent spatial second-order differential operations to extract curvature features. This results in numerous spurious interference peaks that mask the true interlaminar shear deformation signal. Therefore, in this embodiment, the spatial signal reconstruction module 302 is configured to process the original strain data using the Savitzky-Golay (SG) filtering algorithm. Unlike traditional moving average filtering, Savitzky-Golay filtering weights the data points within a sliding window, making it equivalent to performing a high-order polynomial least-squares fit on the data within the sliding window. This processing method can smooth out high-frequency random noise while preserving the original width, height, and inflection point features of the signal to the maximum extent, thereby ensuring high fidelity in retaining second-order derivative (curvature) information in the reconstructed data.

[0039] In the specific implementation of step S200, the spatial signal reconstruction module 302 performs the following steps for each sampling time. Along the spatial coordinate axes A sliding window convolution operation is performed on the original strain data. To ensure that the filtering effect is adapted to the deformation characteristics of semi-coal-rock roadways, the key control parameters are limited as follows in this embodiment: For the order of the fitted polynomial Considering that subsequent steps in this embodiment of the invention require extracting local bending features (i.e., second derivative information) reflecting interlayer slippage, the fitting order is... The value should be set to 2 or 3. If the order is too low (e.g., ...), This will cause the data to be linearized, thus losing curvature information; if the order is too high, it may introduce the Runge phenomenon, leading to edge oscillations.

[0040] The window length is set by the window's... It consists of 10 sampling points, and its corresponding physical length is 1000. This physical length The window length needs to cover the expected shear influence domain between coal and rock layers, while avoiding excessive length that would reduce spatial resolution. The shear influence domain between coal and rock layers refers to the spatial distribution range in which the sensing fiber exhibits a distinct "S"-shaped bend or stress concentration characteristics along the borehole axis due to the continuity of the surrounding rock medium and the coordinated deformation of the optical fiber when relative displacement occurs at the coal-rock interface. If the window length is smaller than this influence domain, the filtering process may fragment the complete deformation features; if it is much larger, it may over-smooth local features. In this embodiment, the physical length of the window... The preferred setting is between 0.5 meters and 2.0 meters. The spatial signal reconstruction module 302 uses the actual spatial sampling interval of the fiber optic demodulation device 20. Through formula The specific integer value of the window half-width is calculated. .

[0041] After determining the above parameters, the spatial signal reconstruction module 302 reconstructs the signal for the location in space. The sampling points were selected before and after them. The original data points are weighted and summed to generate reconstructed strain data. The reconstruction calculation formula is as follows: ; in, Indicates at time ,Location The axial strain value after reconstruction by SG filtering; This represents the half-width of the sliding window, that is, the area within the sliding window including the center point. The total number of data points included is Its specific value is determined by the aforementioned physical length. Spatial sampling interval The constraint is set to an integer between 5 and 25; Represents the relative position index within the sliding window, with values ​​ranging from... arrive ; Indicates Within the central window Original strain data corresponding to each location; This represents the Savitzky-Golay convolution coefficients, which are based on a predefined fitting order. With half the width of the window The fixed weight vector is obtained by solving the normal equations using the least squares method. In this embodiment, for a given... and ,coefficient It is a pre-calculated and stored sequence of constants in the system that satisfies the normalization condition. .

[0042] After processing in step S200 above, the spatial signal reconstruction module 302 outputs reconstructed strain data. The data possesses good spatial second-order differentiability in mathematical properties and retains the local abrupt change characteristics reflecting the interlayer shear behavior of coal and rock, providing high signal-to-noise ratio basic input data for gradient calculation in step S300 and curvature extraction in step S400.

[0043] Step S300 is the dynamic calibration stage of the lithological interface, which is mainly performed by the dynamic calibration module 303. This occurs during the initialization phase of the monitoring system (i.e., the monitoring time). To address the issue of unclear coal-rock boundary locations due to errors in geological data or construction deviations, this embodiment utilizes the strain step characteristic induced by the significant difference in elastic modulus between coal and rock masses to achieve automatic inversion and location of the coal-rock interface. Based on this, geological stiffness parameters are established to provide a physical benchmark for subsequent risk-weighted assessment. Specifically, because coal seams are relatively soft (low modulus) while rock strata are relatively hard (high modulus), a significant strain localization phenomenon occurs at the interface under geostress. This physical phenomenon manifests as an extreme value of the strain gradient in the spatial domain, which is the theoretical basis for interface identification in this step.

[0044] In the specific implementation of step S300, the lithological interface dynamic calibration module 303 first calculates the spatial strain gradient distribution. At the initial moment... The surrounding rock is in a relatively static state of tectonic stress equilibrium. The lithological interface dynamic calibration module 303 uses the reconstructed strain data output from step S200. The strain gradient along the borehole depth direction is calculated using the central difference algorithm. The calculation formula is as follows: ; in, Indicates depth The absolute value of the spatial strain gradient at a given point is, in physical terms, the rate of change of strain per unit length. and They represent the initial time. Located in Reconstructed strain values ​​at adjacent sampling points; The spatial sampling interval is the same parameter defined in step S100.

[0045] Subsequently, the coal-rock boundary depth is identified. The lithological interface dynamic calibration module 303 applies the calculated strain gradient sequence. A full-length scan is performed, and local maxima are found based on edge detection principles. In this embodiment, the system uses an adaptive threshold method to determine valid geological boundaries. The specific determination logic is as follows: First, the gradient sequence is calculated. Background noise level First, take the median of the sequence or the arithmetic mean after removing the largest 5% of data; second, set an effective threshold for the gradient. In this embodiment, Set as , where the coefficient The optimal value range is 3 to 5 to ensure the ability to distinguish between real stratigraphic interfaces and random geological disturbances; finally, all values ​​that meet the criteria are selected. The peak location. If multiple distinct peaks exist, the system reads the reference boundary depth from the preset geological exploration data. Calculate the position of each peak and The Euclidean distance is used to determine the final coal-rock boundary depth, with the peak position having the smallest distance selected. .

[0046] Finally, the initial stiffness ratio is calculated. This is done after determining the coal-rock boundary depth. Subsequently, the lithological interface dynamic calibration module 303 spatially divides the entire monitoring area into two sub-domains based on this depth. To automatically distinguish between the coal seam domain and the rock seam domain, and to avoid manual setting errors caused by different drilling directions, the system calculates... The average strain values ​​of the two subdomains. Based on the physical fact that the elastic modulus of coal is less than that of rock mass, under the same stress environment, the system identifies the side with the larger average strain value as the coal subdomain. The side with the smaller average strain value is identified as the rock mass domain. .

[0047] After dividing the region, the system calculates the ratio of the average strain in the two regions, which is defined as the initial stiffness ratio. This parameter reflects the deformation sensitivity of the coal body relative to the rock mass under current geological conditions, and its calculation formula is as follows: ; in, This represents the initial stiffness ratio. This value is dimensionless, and according to the automatic judgment logic described above, its value is always greater than or equal to 1. This represents the average axial strain value determined to be within the coal seam region. Indicates that it has been identified as a rock mass domain The average axial strain value within; Represents coal body domain Total number of data points within; Indicates rock mass domain Total number of data points within; This represents the reconstructed strain value at each location point at the initial moment.

[0048] Through the processing of step S300 above, the system not only automatically locks the key monitoring locations. It also obtained physical parameters characterizing geological conditions. This parameter will be used in subsequent step S600 to adjust the weights of the risk assessment model, enabling the monitoring system to adapt to semi-coal and rock roadway environments with varying degrees of hardness, thus avoiding the problems of missed or false alarms caused by using a fixed threshold.

[0049] Step S400 is the spatial shear feature extraction stage, which is mainly performed by the spatial shear feature extraction module 304. During real-time monitoring, in order to identify the local deformation characteristics induced by interlayer slippage in the semi-coal-rock roadway, this embodiment employs spatial modal analysis. The physical principle is that when shear slippage occurs at the coal-rock interface, the sensing optical fiber passing through this interface is forced to undergo local geometric deformation, exhibiting an "S"-shaped bending state. Unlike conventional roadway convergence deformation (which mainly causes axial tension or compression of the optical fiber), this shear bending may not be significant in terms of strain magnitude, but it exhibits extremely high curvature characteristics in the spatial distribution of strain. Therefore, this step, by calculating the spatial second derivative (i.e., curvature) of the strain data, can effectively filter out axial tension and compression interference and highlight the interlayer slippage signal.

[0050] In the specific implementation of step S400, the spatial shear feature extraction module 304 first calculates the spatial strain curvature. The spatial shear feature extraction module 304 receives the real-time reconstructed strain data output from step S200. For each monitoring time point The system focuses on the coal-rock boundary depth determined in step S300. The spatial strain curvature of this region and its neighborhood are calculated using a second-order central difference algorithm. This parameter characterizes the degree of bending of the optical fiber along its axial direction, and its calculation formula is as follows: ; in, Indicates at time ,depth The spatial strain curvature at a given point, its physical unit is m. -2 ; , and These represent the reconstructed strain values ​​at the current position and the adjacent sampling points before and after it, respectively. The spatial sampling interval is determined by the hardware configuration of the fiber demodulation device 20 and is used here as the step size for the differential operation.

[0051] Subsequently, the shear decoupling index is calculated. To eliminate the interference of inherent system noise (such as temperature fluctuations and equipment thermal noise) on the monitoring results and to provide a dimensionless quantitative index, the spatial shear feature extraction module 304 introduces the signal-to-noise ratio (SNR) analysis concept and normalizes it using the standard deviation of rock mass background noise. In this embodiment, based on the characteristics of high elastic modulus and high deformation stability of rock strata, the system selects the rock mass domain. As a relatively stable geological reference zone, the standard deviation of the curvature distribution within this region is calculated. Subsequently, the depth of the coal-rock boundary was calculated. The ratio of the peak curvature at a given point to the standard deviation of the background noise is defined as the shear decoupling index. The calculation formula is as follows: ; in, Indicates time The shear decoupling index is a dimensionless value that reflects the degree of significance of bending deformation at the interface relative to background noise. Indicates the depth of the coal-rock boundary The absolute value of the spatial strain curvature at that location; Indicates time Standard deviation of curvature background noise within the rock mass domain; Indicates rock mass domain The number of sampling points within; Indicates time The arithmetic mean of the curvature of all sampling points within the rock mass domain.

[0052] Finally, spatial threshold discrimination is performed. The spatial shearing feature extraction module 304 is configured with a preset spatial threshold. This spatial threshold is based on statistical methods. Criteria are set to determine whether the current bending characteristic constitutes a statistically significant anomalous event. In this embodiment, The preferred value range is 3 to 6, that is, when the curvature amplitude at the interface exceeds 3 to 6 times the standard deviation of the background noise level, it is determined that a non-random structural deformation has occurred.

[0053] The system will calculate the shear decoupling index in real time. Spatial threshold Compare. If This indicates the presence of significant spatial shearing at the coal-rock interface, and the spatial shearing feature extraction module 304 immediately generates a trigger signal. This trigger signal is used to activate the subsequent time-domain rheological feature extraction module 305 for further time-dimensional verification; if This indicates that the current state is within the background fluctuation range, and the system can directly determine that the current state is safe, or the rheological decoupling degree can be adjusted. The value is assigned to zero, and the process proceeds to step S600 for verification and recording of the comprehensive risk index to ensure that the risk status throughout the entire cycle is quantitatively assessed.

[0054] Step S500 is the time-domain rheological feature extraction stage, which is mainly performed by the time-domain rheological feature extraction module 305. When the shear decoupling index in step S400... When the spatial threshold is exceeded, it indicates that the monitoring system has detected suspected interlayer slippage characteristics in the spatial dimension. To further confirm whether this characteristic originates from the continuous rheological failure of the surrounding rock and to rule out the interference of transient vibrations (such as blasting vibrations and mechanical disturbances), this embodiment introduces long-period time window analysis technology. Its physical principle is based on the theory of rock rheology: the instability and failure of the surrounding rock usually goes through decelerating creep (first stage), steady-state creep (second stage), and accelerated creep (third stage). Although transient disturbances can cause abrupt strain changes, they lack sustained acceleration characteristics; while the occurrence of interlayer slippage is often accompanied by the coal and rock medium entering the accelerated creep stage. Therefore, this step quantifies the non-cooperation of the two-phase medium in the dynamic evolution trend by calculating the second derivative of strain with time (strain acceleration), thereby accurately identifying the pre-disaster state.

[0055] In the specific implementation of step S500, the time-domain rheological feature extraction module 305 first constructs a rheological feature time window. This takes into account the sampling frequency of conventional fiber optic demodulation devices. Typically, the Hz levels are high (e.g., 1 Hz to 10 Hz). If strain acceleration is calculated directly using adjacent sampling points (with time intervals of only 0.1 to 1 second), the differential results will be overwhelmed by noise or approach zero because the changes in geological creep on a microsecond-scale timescale are much smaller than the system measurement noise. Therefore, this embodiment defines a rheological characteristic time window. This parameter, serving as the time step for sparsification, is used to extract long-period trends from high-frequency data streams. In this embodiment, The value is set based on the theoretical rheological rate of the surrounding rock, with an optimal range of 10 minutes to 6 hours. During data processing, the system extracts the current time. Historical Moments as well as The reconstructed strain data is used in the calculation, thus mathematically guaranteeing the effective number of digits in the difference operation and realizing the effective extraction of low-frequency creep characteristics.

[0056] Subsequently, the time-dependent deformation factor (TDF) is calculated. The system is based on a selected rheological characteristic time window. The strain acceleration distribution within the entire monitoring range is calculated and defined as the strain aging factor. This factor reflects the rate of change of the surrounding rock deformation rate and is a key dynamic indicator for determining whether the rock mass has entered the accelerated failure stage. Using a second-order backward difference algorithm, its calculation formula is as follows: ; in, Indicates at time ,depth The strain aging factor (i.e., strain acceleration) at a given location is expressed in units of . (See (Depending on the time unit) This represents the reconstructed strain value at the current moment; This represents the historical reconstructed strain value prior to a time window; This represents the historical reconstructed strain values ​​two time windows prior; The length of the rheological characteristic time window needs to be converted to a standard time unit when used in the formula.

[0057] Finally, the rheological decoupling degree is calculated. To quantify the differences in temporal evolution between the two sides of the coal-rock interface, the time-domain rheological feature extraction module 305 combines the coal body domains divided in step S300. With rock mass domain The absolute value of the difference between the average strain accelerations in the two regions is calculated and defined as the rheological decoupling degree. The physical significance of this parameter lies in the fact that, during normal coordinated settlement, although the deformation rates of coal and rock differ, their acceleration trends should remain consistent (i.e., both decelerating or in a steady state). However, when interlayer displacement occurs, weaker coal seams often enter an accelerated creep state first, while harder rock layers may remain in a steady state. This inconsistency in dynamic pace leads to… The value increases significantly. The calculation formula is as follows: ; in, Indicates time The rheological decoupling degree; the larger the value, the stronger the dynamic separation tendency of the coal and rock two-phase media. This represents the average strain acceleration within the coal seam region; This represents the average strain acceleration within the rock mass domain; and These represent the number of spatial sampling points contained within the coal mass domain and the rock mass domain, respectively, defined in step S300; The strain aging factor at each point is calculated in step S500.

[0058] Through the processing in step S500 above, the system extracts quantitative indicators characterizing the dynamic trend of inter-layer slippage from the time dimension. This indicator effectively addresses the limitation of relying solely on spatial morphology (step S400) to determine the stage of deformation development, providing a key criterion for subsequent multidimensional risk assessment regarding whether deformation is accelerating.

[0059] Step S600 is the multidimensional coupling criterion and risk assessment stage, which is mainly executed by the multidimensional coupling criterion module 306. This is after obtaining the initial stiffness ratio. (From step S300), shear decoupling index (From step S400) and rheological decoupling (From step S500) Following this, to address the issue of false alarms or missed alarms caused by single indicators in complex geological environments, this embodiment constructs a comprehensive risk assessment model that integrates geological attributes, spatial morphology, and temporal evolution trends. This embodiment employs a geological attribute-based assessment strategy: the stability of the tunnel roof depends not only on the current monitoring data volume but also on the physical properties of the surrounding rock itself. For tunnels with large stiffness differences (i.e.,...) Large (soft and hard interlayered structures) have a weaker ability to resist shear failure, and therefore should be judged to have a higher risk level under the same deformation characteristics.

[0060] In the specific implementation of step S600, the multidimensional coupling criterion module 306 first performs parameter normalization and dimensional unification. Due to the shear decoupling index... It is a dimensionless ratio, while the rheological decoupling degree Having physical units (e.g.) Since the two cannot be directly algebraically operated on, this embodiment first introduces a critical rheological threshold. right Standardize the process. The critical acceleration value characterizing the coal and rock mass entering the irreversible accelerated failure stage (the third creep stage). In specific implementation, The threshold value can be obtained by conducting graded loading rheological experiments on coal-rock composite specimens in the field; if experimental conditions are lacking, based on conventional coal-rock mechanical parameters in this field, the threshold value is preferably set to 10 to 50. The system calculates standardized rheological indices. This transforms the features of the time dimension into those of the spatial dimension. Dimensionless values ​​of the same order of magnitude.

[0061] Subsequently, a multidimensional coupled risk index model is constructed. This embodiment uses a weighted Euclidean distance model to calculate the comprehensive risk index. And introduce an initial stiffness ratio As a nonlinear amplification factor, this model is used to characterize the degree to which the current monitoring state deviates from the safety origin in the two-dimensional risk plane of "space-time". Its calculation formula is as follows: ; in, Indicates time The comprehensive risk index is a dimensionless scalar that characterizes the degree of risk of interlayer slippage in the roadway roof. The initial stiffness ratio calculated in step S300 is used as a geological sensitivity coefficient in the calculation to amplify the risk value in a soft stratum environment. This is a geological sensitivity index used to adjust the impact of the initial stiffness ratio on the overall risk. Considering that the failure of coal-rock assemblies often exhibits nonlinear characteristics, this embodiment... The preferred value range is 0.5 to 1.0; The shear decoupling index is calculated in step S400; The spatial threshold set in step S400 is used here as the normalization benchmark for the spatial dimension; The rheological decoupling degree calculated in step S500; The critical rheological threshold determined in step S600; and These are the spatial feature weights and the temporal feature weights, respectively, and satisfy the following conditions: In this embodiment, based on the evolutionary pattern that inter-layer faulting typically begins with localized spatial bending (precursor) followed by macroscopic acceleration (damage) over time, the system sets a spatial weight slightly greater than a temporal weight to improve early warning capabilities. The preferred configuration is as follows: .

[0062] Finally, the risk level is determined and control commands are output. The multi-dimensional coupling criterion module 306 is configured with hierarchical early warning logic, which calculates the risk level. The current roadway safety status is determined by comparing the current status with a preset risk classification threshold. The specific classification criteria are as follows: A safe state, that is, when When the monitoring area is determined to be within the range of elastic deformation or stable creep, the system maintains the default low-frequency sampling mode. The low-frequency sampling mode means that the system collects data according to the preset reference frequency (preferably 0.1Hz to 1.0Hz). This frequency is sufficient to cover the slow background creep characteristics of the surrounding rock, and can effectively reduce data redundancy and storage load. Attention status, that is, when When a statistical anomaly is detected in the monitoring data, the system automatically switches to a high-frequency sampling mode for encrypted monitoring. Specifically, the system increases the sampling frequency to the high-frequency range allowed by the device (e.g., 5.0Hz to 10.0Hz) to capture possible transient dynamic precursors. Warning status, that is, when When the system determines that there is a clear shear slip trend at the coal-rock interface and that the geological sensitivity is high, it generates a yellow warning signal and triggers remote data push. Alarm status, i.e. when If a high risk of inter-layer slippage and instability is detected, the system generates a red alarm signal and activates the underground audible and visual alarm device to prompt personnel to evacuate.

[0063] Through the processing in step S600 above, this embodiment of the invention achieves an upgrade from monitoring a single physical quantity to multi-dimensional state assessment. Specifically, by introducing the geological stiffness ratio... As an amplification factor, the monitoring system becomes more sensitive to high-risk areas such as soft rock or fractured zones, thereby effectively improving the accuracy and pertinence of roof disaster early warning in semi-coal and rock roadways.

[0064] To further clarify the collaborative working process of the technical solution of this invention, a specific working scenario example will be used below. Assume the monitoring object is a typical deep semi-coal-rock roadway, and the distributed fiber optic sensing device 10 is deployed in a roof borehole at a depth of 30 meters. The spatial resolution of the fiber optic demodulation device 20 is set to 0.2 meters, and the time sampling frequency is set to 1 Hz. The various modules in the data processing terminal 30 operate collaboratively according to the aforementioned process.

[0065] During the initialization phase of the monitoring system, the data acquisition module 301 acquires the raw strain data at the initial moment. At this time, the data contains slight random fluctuations due to thermal noise from the photoelectric devices. The spatial signal reconstruction module 302 receives this data and applies the Savitzky-Golay filtering algorithm for smoothing, effectively filtering out high-frequency noise. Following this, the lithological interface dynamic calibration module 303 performs gradient calculations on the reconstructed spatial strain distribution. (See reference...) Figure 3 , Figure 3 This embodiment demonstrates the strain distribution along the borehole depth and the corresponding curvature characteristic curve at a key monitoring moment. Figure 3 In the diagram, the horizontal axis represents the borehole depth, the left vertical axis represents the reconstructed axial strain, and the right vertical axis represents the spatial strain curvature. During system initialization (corresponding to the background baseline in the diagram), the lithological interface dynamic calibration module 303 identified a significant strain gradient maxima at a depth of 18.5 meters, thus automatically determining this location as the coal-rock boundary depth. Simultaneously, the lithological interface dynamic calibration module 303 calculates the average strain ratio on both sides of the boundary depth and determines the initial stiffness ratio. The value of 2.5 indicates that the coal seams in this area are relatively soft and have high geological sensitivity.

[0066] As monitoring progressed, the surrounding rock of the tunnel began to deform under the influence of mining stress. Around the 50th hour of monitoring, initial interlayer displacement occurred within the surrounding rock. At this time, data acquisition module 301 continuously collected data, and spatial signal reconstruction module 302 output reconstructed strain data in real time. Spatial shear feature extraction module 304 captured the strain curve near the boundary depth of 18.5 meters, showing... Figure 3The "S"-shaped bending feature is shown. Correspondingly, the calculated spatial strain curvature shows a sharp peak at 18.5 meters. The spatial shear feature extraction module 304 calculates the shear decoupling index at this point. The value was found to have climbed to 4.2, exceeding the preset spatial threshold. (Set to 3.5).

[0067] In response to the trigger signal issued by the spatial shear feature extraction module 304, the time-domain rheological feature extraction module 305 immediately starts. This module calls historical data from the past 3 hours, sets the rheological time window to 30 minutes, and performs differential calculations on the strain acceleration on both sides of the boundary depth. The calculation results show that the coal body side exhibits obvious accelerated creep characteristics, while the rock mass side remains relatively stable, resulting in rheological decoupling. It grows rapidly.

[0068] The multidimensional coupling criterion module 306 receives the above indicators and combines them with the initial stiffness ratio. Perform a weighted calculation. See also Figure 4 , Figure 4 The graph illustrates the evolution trends of the shear decoupling index, rheological decoupling degree, and comprehensive risk index over monitoring time in this embodiment. Figure 4 In the graph, the horizontal axis represents the monitoring time, and the vertical axis represents the normalized index value. Figure 4 The three curves represent the shear decoupling index. The standardized rheological decoupling degree and the final calculated comprehensive risk index .like Figure 4 As shown, in the initial monitoring period (0-40 hours), all indicators fluctuated at low levels, and the system was in a safe state. Around the 50th hour, the shear decoupling index... First to break through the threshold (corresponding to) Figure 3 (The moment when the spatial curvature occurs is shown). At this point, although the spatial characteristics are obvious, the rheological decoupling degree has not yet fully responded, resulting in a comprehensive risk index. Rising to around 2.5, the multidimensional coupling criterion module 306 determined that the system entered an early warning state, indicating the presence of a shearing trend. Subsequently, between the 55th and 60th hours, with the hysteretic decoupling degree showing a significant increase, it indicated that interlayer faulting had transformed into irreversible accelerated damage, and the comprehensive risk index... Driven by dual factors and amplified by the geological stiffness ratio, the critical value or alarm threshold of 4.0 is rapidly exceeded. At this point, the multi-dimensional coupling criterion module 306 immediately outputs an alarm status command, triggering the audible and visual alarm device, thus achieving precise capture of interlayer shear instability events. As can be seen from the above process, this system utilizes spatial characteristics for sensitive triggering, temporal characteristics for trend confirmation, and geological parameters for risk weighting, effectively avoiding misjudgments based on a single indicator and achieving accurate monitoring across all times and spaces.

[0069] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A real-time monitoring system for surrounding rock in semi-coal and rock roadways based on fiber optic sensing technology, characterized in that, include: The data acquisition module is used to acquire raw strain data from the distributed fiber optic sensing device inside the surrounding rock borehole from the fiber optic demodulation device. The spatial signal reconstruction module is used to perform spatial domain smoothing fitting on the original strain data to output reconstructed strain data. The lithological interface dynamic calibration module is used to identify the coal-rock boundary depth based on the spatial gradient distribution of the reconstructed strain data and to calculate the initial stiffness ratio. The spatial shear feature extraction module is used to calculate the spatial strain curvature at the coal-rock boundary depth based on the reconstructed strain data, and to calculate the shear decoupling index based on the spatial strain curvature. When the shear decoupling index exceeds a spatial threshold, a trigger signal is generated. The time-domain rheological feature extraction module is used to respond to the trigger signal and calculate the strain aging factor on both sides of the coal-rock boundary depth, and calculate the rheological decoupling degree based on the strain aging factor on both sides. The multidimensional coupling criterion module is used to calculate a comprehensive risk index based on the initial stiffness ratio, the shear decoupling index and the rheological decoupling degree, and output a risk warning command based on the comprehensive risk index; The spatial shearing feature extraction module is specifically used for: The second-order central difference algorithm is used to calculate the coal-rock boundary depth and the spatial strain curvature within the neighborhood of the coal-rock boundary depth. A rock mass domain was selected as a geological reference area, and the standard deviation of the background noise of the curvature distribution within the rock mass domain was calculated. Calculate the ratio of the peak curvature at the coal-rock boundary depth to the standard deviation of the background noise, and define the ratio as the shear decoupling index; The shear decoupling index calculated in real time is compared with the spatial threshold. If the shear decoupling index is greater than the spatial threshold, it is determined that spatial shearing signs have appeared at the coal-rock interface. The time-domain rheological feature extraction module is specifically used for: Define a rheological time window as the time step for sparsification, and extract the reconstructed strain data of the current moment, the historical moment before one rheological time window, and the historical moment before two rheological time windows; Using the second-order backward difference algorithm, the strain acceleration distribution within the entire monitoring range is calculated based on the rheological time window, and the strain acceleration distribution is defined as the strain aging factor. Based on the coal body domain and rock body domain divided by the coal-rock boundary depth, the average strain aging factor in the coal body domain and the average strain aging factor in the rock body domain are calculated respectively. Calculate the absolute value of the difference between the average strain aging factor in the coal body domain and the average strain aging factor in the rock body domain, and define the absolute value of the difference as the rheological decoupling degree.

2. The real-time monitoring system for surrounding rock in semi-coal and rock roadways based on fiber optic sensing technology according to claim 1, characterized in that, The data acquisition module is specifically used for: Establish a spatial coordinate axis with the borehole opening as the origin and the direction of borehole depth as the positive direction, and a physical time coordinate axis with the monitoring start time as the starting point; The spatial location of the reflection point is located using optical time-domain reflectometry, and the axial strain value at each point along the optical fiber is calculated by using the linear mapping relationship between the frequency shift characteristic of the backscattered light signal and the axial strain. The calculated axial strain value is mapped to a spatiotemporal coordinate system composed of the spatial coordinate axis and the physical time coordinate axis to generate the original strain data.

3. The real-time monitoring system for surrounding rock in semi-coal and rock roadways based on fiber optic sensing technology according to claim 2, characterized in that, The spatial signal reconstruction module is specifically used for: For each sampling time, the physical length of the sliding window is set to cover the expected inter-layer shear influence domain of coal and rock, and the half-width of the sliding window is calculated based on the spatial sampling interval. A sliding window convolution operation is performed on the original strain data along the spatial coordinate axis. For the sampling point located at the center point of the sliding window, the original data points before and after the center point are selected and weighted and summed to obtain the single-point reconstructed strain value corresponding to the center point. The reconstructed strain data is composed of the single-point reconstructed strain values ​​corresponding to all sampling positions and is then output.

4. The real-time monitoring system for surrounding rock in semi-coal and rock roadways based on fiber optic sensing technology according to claim 1, characterized in that, The dynamic calibration module for lithological interfaces is specifically used for: The spatial strain gradient along the borehole depth direction of the reconstructed strain data is calculated using the central difference algorithm. The spatial strain gradient is scanned along its entire length to identify local maxima. The peak position with the smallest distance is selected as the final coal-rock boundary depth in combination with a preset reference boundary depth. Based on the coal-rock boundary depth, the monitoring area is spatially divided into a first spatial subdomain and a second spatial subdomain, and the average strain value in each spatial subdomain is calculated. The ratio of the average strain value in the first spatial subdomain to that in the second spatial subdomain is defined as the initial stiffness ratio.

5. The real-time monitoring system for surrounding rock in semi-coal and rock roadways based on fiber optic sensing technology according to claim 1, characterized in that, The multidimensional coupling criterion module is specifically used for: A critical rheological threshold is introduced to standardize the rheological decoupling degree, and standardized rheological indices are calculated. The initial stiffness ratio is adjusted by introducing a geological sensitivity index, and the nonlinear amplification factor is calculated. A weighted Euclidean distance model is constructed, and the weighted sum of squares is calculated for the ratio of the shear decoupling index to the spatial threshold and the ratio of the rheological decoupling degree to the critical rheological threshold, respectively. The weighted sum of squares is square-rooted, and the result is multiplied by the nonlinear amplification factor to obtain the comprehensive risk index.

6. The real-time monitoring system for surrounding rock in semi-coal and rock roadways based on fiber optic sensing technology according to claim 4, characterized in that, The lithological interface dynamic calibration module, after spatially dividing the monitoring area into a first spatial subdomain and a second spatial subdomain based on the coal-rock boundary depth, is also used for: Based on the physical property that the elastic modulus of coal is less than that of rock mass, the average strain values ​​of the first spatial subdomain and the second spatial subdomain are compared. The spatial subdomain on the side with the larger average strain value is identified as the coal mass domain, and the spatial subdomain on the side with the smaller average strain value is identified as the rock mass domain.

7. The real-time monitoring system for surrounding rock in semi-coal and rock roadways based on fiber optic sensing technology according to claim 5, characterized in that, In the weighted Euclidean distance model, spatial feature weights and temporal feature weights are set, and the sum of the spatial feature weights and the temporal feature weights is 1, wherein the spatial feature weights are greater than the temporal feature weights.

8. The real-time monitoring system for surrounding rock in semi-coal and rock roadways based on fiber optic sensing technology according to claim 1, characterized in that, When the multidimensional coupling criterion module outputs a risk warning command, it specifically executes the following logic: The calculated comprehensive risk index is compared with a preset risk classification threshold to determine the risk status; Based on the determined risk status, the monitoring parameters are automatically adjusted and / or corresponding early warning signals are output.

Images