A coal mine floor water guide channel dynamic identification and early warning method based on microseism-electric coupling

Abstract

This invention provides a dynamic identification and early warning method for water-conducting channels in coal mine floors based on microseismic-electrodynamic coupling. The method includes: acquiring multi-dimensional basic data of the target working face and affected area in the mine to construct an initial floor water-conducting risk model; designing a monitoring network model by combining microseismic and electrodynamic data; collecting microseismic spatiotemporal response data to obtain a set of floor crack evolution characteristics; triggering zonal electrodynamic excitation and response acquisition to obtain a set of water-bearing crack electrodynamic response characteristics; constructing a spatiotemporal coupling field combining microseismic and electrodynamic data to obtain a set of candidate water-conducting channels; determining whether the water-conducting channel is connected; if so, calculating the water-conducting capacity to obtain a channel water-conducting capacity evaluation set; classifying risk levels and implementing corresponding early warnings and differentiated management for different risk levels. This invention can also achieve integrated intelligent monitoring of the health status assessment, early degradation warning, and dynamic prediction of remaining life of key components of wind turbines.

Description

Technical Field

[0001] This invention relates to the field of coal mine floor water channel identification technology, and in particular to a dynamic identification and early warning method for coal mine floor water channels based on micro-vibration-electric coupling. Background Technology

[0002] Coal mine floor water-conducting channels refer to seepage paths within the coal seam floor strata, such as fissures, faults, collapse columns, karst fissure zones, and mining-induced damage zones, that allow confined water, karst water, or groundwater from aquifers to migrate into the mining space. Coal seam mining continuously alters the stress, fissures, and permeability of the floor rock mass. Structures that were originally non-conductive or had weak water-conductivity may expand, connect, and form water inrush channels due to mining activities. Therefore, identifying and providing early warning of coal mine floor water-conducting channels is of paramount importance.

[0003] The existing mainstream identification methods fall into three main categories: static advanced detection technology, downhole dynamic real-time monitoring technology, and multi-field coupling fusion technology, covering the entire process of pre-mining exploration and dynamic tracking throughout the entire mining process. However, the existing methods and technologies have the following shortcomings: 1) Separation of environment and operating conditions: Traditional condition monitoring only focuses on equipment vibration, temperature and other response signals, ignoring the coupling effect between the engine room environment and operating conditions; 2) Ignoring scenario differences: The operating scenarios are not identified, making it impossible to distinguish between normal performance under harsh conditions and abnormal performance under mild conditions; 3) Independent evaluation of single components: Each component is evaluated in isolation, lacking overall machine collaborative analysis and consideration of coupling risk transmission; 4) Delayed detection of early degradation: Alarms are only triggered when the health index declines, lacking more sensitive early degradation indicators; 5) Coarse prediction of remaining life: Simple linear extrapolation or fixed models are used, without considering the accelerated impact of environmental and load shocks on lifespan. Summary of the Invention

[0004] To overcome the shortcomings of existing technologies, the purpose of this invention is to provide a dynamic identification and early warning method for water diversion channels in coal mine floor based on microseismic-electric coupling. By extracting coupling features and identifying scenarios of the nacelle environment and operating conditions, a dynamic health benchmark and a multi-dimensional residual analysis system are constructed to achieve integrated intelligent monitoring of the health status assessment, early degradation warning and dynamic prediction of remaining life of key components of wind turbines.

[0005] To achieve the above objectives, the present invention provides the following solution: a method for dynamic identification and early warning of water diversion channels in coal mine floor based on microseismic-electric coupling, comprising: Acquire multidimensional basic data of the target working face and the affected area of ​​the mine, and use the multidimensional basic data to construct an initial floor water-conducting risk model; Based on the initial bottom plate water-conducting risk model, and combined with microseismic data and electrodynamic data, a monitoring network model including monitoring zones and an adaptive monitoring mechanism is designed. The monitoring network model is used to collect microseismic spatiotemporal response data to obtain a set of characteristics of the evolution of base plate cracks; Based on the aforementioned base plate fracture evolution feature set, regional electric excitation and response acquisition are triggered to obtain the water-bearing fracture electric response feature set; Using the base plate crack evolution feature set and the water-bearing crack electrodynamic response feature set, a spatiotemporal coupled field combining microseismic and electrodynamic data is constructed to obtain a candidate water-conducting channel set; Using the candidate water diversion channels and hydrological dynamic data, it is determined whether the water diversion channels are connected. If so, the water diversion capacity is calculated to obtain the channel water diversion capacity evaluation set. The risk level is classified using the water conduction capacity evaluation set, and corresponding early warning and differentiated management are carried out for different risk levels.

[0006] Optionally, acquire multi-dimensional basic data of the target working face and affected area of ​​the mine, and use the multi-dimensional basic data to construct an initial floor water-conducting risk model, including: Collect geological structure data, hydrogeological data, mining engineering data, rock physics and geostress data, mining-induced damage prediction data, and existing geophysical and drilling anomaly data of the target working face and its affected area to obtain multidimensional raw data; The multidimensional raw data are uniformly entered into the geological modeling database, and spatial coordinate transformation is performed in the local three-dimensional rectangular coordinate system with the starting point of the target working face as the origin to obtain multidimensional basic data. Based on the aforementioned multidimensional basic data, 3D geological modeling software is used to construct stratigraphic interface elements, structural entity elements, hydrological attribute elements, and mining space elements to obtain an initial base plate model. Based on the initial base plate model, the study area is discretized into three-dimensional regular hexahedral grid cells, and a three-dimensional index and centroid coordinates are set for each grid cell; For each grid cell, quantify the lithology and integrity parameters, structural development parameters, hydraulic distance parameters from the aquifer, equivalent hydrostatic pressure influence coefficient parameters, initial permeability parameters, and mining-induced damage influence parameters to obtain the initial properties of each grid cell and output the initial floor water conduction risk model.

[0007] Optionally, based on the aforementioned multidimensional basic data, three-dimensional geological modeling software is used to construct stratigraphic interface elements, structural entity elements, hydrological attribute elements, and mining space elements to obtain an initial base model, including: Based on the aforementioned multidimensional basic data, using three-dimensional geological modeling software, the coal seam floor interface, the top and bottom interfaces of each aquifer and aquifer are generated sequentially, and the fault plane or column is used to perform three-dimensional characterization of the fault and collapse column to obtain the stratigraphic interface elements. Faults are interpreted as three-dimensional surface or volumetric features of attitude, width, and conductivity, and collapse columns are interpreted as columnar anomalies, thus obtaining structural entity features. Aquifer water pressure, permeability coefficient and aquitard integrity are treated as continuous attribute fields and interpolated to the entire model space using Kriging or sequential Gaussian simulations to obtain hydrological attribute volume elements. Construct a cavity model of the working face cut, two roadways, and planned advance range, and mark adjacent goaf areas to obtain mining space elements; By combining the stratigraphic interface elements, the structural entity elements, the hydrological attribute elements, and the mining space elements, an initial base plate model is obtained.

[0008] Optionally, based on the initial bottom plate water-conducting risk model, and combining microseismic data and electrodynamic data, a monitoring network model including monitoring zones and an adaptive monitoring mechanism is designed, including: Based on the initial attributes of each grid cell in the initial bottom plate water-conducting risk model, the parameters are weighted and fused to obtain a comprehensive risk score. According to the comprehensive risk score, the study area is divided into a red key monitoring area, a yellow attention monitoring area, and a green background monitoring area. Based on the divided monitoring areas, microseismic monitoring sensors, electric detection sensors, and hydrological dynamic monitoring sensors with different densities are placed in different areas to obtain an initial sensor deployment scheme. Based on the initial sensor deployment scheme, a forward-moving strategy triggered by mining advancement and a high-density supplementation strategy triggered by anomalies are designed to obtain an adaptive monitoring mechanism and output a monitoring network model. The microseismic monitoring sensor includes a downhole three-component microseismic detector, a single-component high-frequency microseismic detector, and an acceleration sensor; the electric detection sensor includes a power supply electrode, a measuring electrode, and a spontaneous potential sensor; and the hydrological dynamic monitoring sensor includes a stratified water pressure sensor, a stratified flow sensor, a hydrological sensor, and an online hydrochemical sensor.

[0009] Optionally, the forward-moving strategy includes: defining a monitoring window length starting from the opening eye in the working face advancing direction; the window automatically moves forward with each step the working face advances, and triggers a sensor switching action. The high-density replenishment strategy includes: by setting an abnormal threshold, when the energy release rate or frequency of microseismic events in a certain area suddenly increases, or the apparent resistivity decreases by more than 10%, or the natural potential shows a continuous drift of more than 5mV, the area is automatically labeled as a risk activation zone and the backup sensor activation action is triggered.

[0010] Optionally, the monitoring network model is used to collect microseismic spatiotemporal response data to obtain a set of base plate crack evolution characteristics, including: The monitoring network model is used to collect continuous data. The collected continuous data is automatically denoised, and waveform segments of suspected microseismic events are extracted. Longitudinal and transverse waves are extracted from the waveform segments of suspected microseismic events to obtain sub-millisecond microseismic events. Based on the microseismic events, the location and time of rupture occurrence are inverted using the initial bottom plate water-conducting risk model to obtain the rupture source of the microseismic events. Then, for the microseismic events, the microseismic energy, dominant frequency and spectral characteristics, moment tensor inversion and rupture type, and main direction of crack propagation are calculated to obtain the microseismic event archive. Based on the microseismic events, a moving time window is defined. Within the moving time window, it is identified whether there is a sudden increase in event frequency, a continuous increase in event energy, a movement of the average event depth towards the deeper aquifer, an directional arrangement of events along known faults in the initial bottom plate water-conducting risk model, a sudden increase in the proportion of tensional events, a significant increase in the proportion of low-frequency events, and a continuous expansion of the event spatial volume, thereby obtaining abnormal signals. Based on the anomalous signals, risk quantification is performed to calculate the crack activity score and obtain the base plate crack evolution index. Then, the microseismic time archive and the base plate crack evolution index are integrated to obtain the base plate crack evolution feature set.

[0011] Optionally, based on the base plate fracture evolution feature set, zonal electrodynamic excitation and response acquisition are triggered to obtain the water-bearing fracture electrodynamic response feature set, including: For the aforementioned base plate crack evolution index, a focus threshold is set to determine whether the base plate crack evolution index of each grid cell exceeds the focus threshold. If so, it is determined that shear fracture, tension fracture, or crack is extending into the aquifer and triggering the physical examination threshold. The area to be examined is then delineated by expanding outward from the grid cell as the center. Based on the area to be inspected, the base plate crack evolution index is used to perform low-frequency pulse electric field excitation, multi-frequency sweep frequency excitation, excitation polarization attenuation measurement, natural potential dynamic monitoring and zoned cross power supply measurement to complete multi-mode joint electric excitation; Based on the data collected by the multi-mode joint electric excitation, inversion is performed to calculate the real-time resistivity, polarizability, excitation polarization decay time constant, pulse electric field response amplitude, natural potential change, and electric response phase difference, thereby obtaining the real-time electric response parameters. The real-time electrodynamic response parameters are compared with the original electrical background values ​​in the initial bottom plate water-conducting risk model to calculate the resistivity change and spontaneous potential change. The changes are weighted and fused to obtain the hydrodynamic activation index of the aquifer. Based on the hydrodynamic activation index, the hydrodynamic response feature set of the aquifer in the study area is output. The hydrodynamic response feature set of the aquifer includes resistivity change, excitation polarization characteristics of the aquifer, directional change of spontaneous potential, and impulse response.

[0012] Optionally, using the base plate fracture evolution feature set and the water-bearing fracture electrodynamic response feature set, a spatiotemporal coupled field combining microseismic and electrodynamic data is constructed to obtain a candidate water-conducting channel set, including: The evolution feature set of the base plate cracks and the electrodynamic response feature set of the water-bearing cracks are mapped to the initial base plate water-conducting risk model to unify the spatiotemporal coordinate system. In each grid cell, the microseismic data and electrodynamic data are compared to calculate the spatial overlap of the anomalies and the time difference of the anomaly occurrence. Determine whether the spatial overlap and time difference of the grid cells meet the preset range standard. If so, according to the main direction of fracture propagation, the direction of electrodynamic anomaly gradient and the direction pointing to the aquifer, the grid cells that meet the connectivity, directional consistency and temporal logic are connected into an ordered set by the shortest path algorithm to obtain the candidate water-conducting channel. The starting and ending points, direction and tendency, extension length, volume and dynamic evolution rate of the candidate water-conducting channels are calculated daily to construct a candidate water-conducting channel profile.

[0013] Optionally, using the candidate water-diverting channels and hydrological dynamic data, it is determined whether the water-diverting channels are connected. If so, the water-diverting capacity is calculated to obtain a channel water-diverting capacity evaluation set, including: For each of the candidate water diversion channels, the water source and outlet are acquired, and all hydrological sensor data in the area where the candidate water diversion channel is located are extracted to calculate water pressure changes, stratified flow changes, and water temperature and water chemistry tracers to obtain hydrological dynamic data. Based on the water source, outlet, and hydrological dynamic data, the probability of the candidate water diversion channel being connected is calculated using the Sigmoid function. According to the magnitude of the connection probability, the channel connection status is divided into: not connected, potentially connected, critically connected, and connected. Based on the channel's connectivity status, an equivalent water-conducting model of the candidate water-conducting channel is constructed using Darcy's law. The equivalent permeability coefficient is then inverted using soft computing, the equivalent water-passing area is calculated, the water pressure difference and length are obtained, and the water-conducting direction and potential water inrush location are determined to form a water-conducting capacity evaluation set for the water-conducting channel. This evaluation set includes the channel ID and spatial location, geometric dimensions, fracture development degree, water filling degree, connectivity probability, equivalent water-conducting capacity, potential water inrush location, and risk development rate.

[0014] Optionally, the water-conducting capacity evaluation set is used to classify risk levels, and corresponding early warnings and differentiated management are carried out for different risk levels, including: Based on the water-conducting capacity evaluation set, parameter weighting and fusion are performed to obtain a risk index. Based on the risk index, the risk level is divided into blue warnings (indicating overall calm bottom), yellow warnings (indicating signs of activity), orange warnings (indicating clear danger and risk of deterioration), and red warnings (indicating a high probability of water inrush). When the risk level is determined to be a blue alert, no risk measures will be taken; When the risk level is determined to be a yellow alert, the frequency of microseismic and electrodynamic data acquisition is increased, and the specific locations that need to be verified by intensive drilling are sent to remind the team to execute within a time limit. When the risk level is determined to be an orange alert, a speed limit command is sent directly to the coal mining machine and the dispatch center, or it is recommended to partially stop mining and activate the preset water exploration or grouting plan. When the risk level is determined to be a red alert, the underground audible and visual alarm will be triggered immediately, the unnecessary power supply to the dangerous area will be automatically cut off, the evacuation route map will be pushed to all relevant personnel, and emergency response measures will be initiated simultaneously through the ground command center.

[0015] This invention discloses the following technical effects by providing a dynamic identification and early warning method for water guiding channels in coal mine floor based on microseismic-electric coupling: 1. Comprehensive perception of environment, working conditions, and status: covering three levels: cabin microenvironment, operating load, and equipment response, breaking the limitation of traditional systems that only focus on a single physical quantity.

[0016] 2. Coupled Feature-Driven Scene Recognition: By combining the environment and operating conditions and using scene tags, the limitations of traditional environmental monitoring and operating condition monitoring are overcome. The combined effects of environmental stress and operating load are explicitly modeled, enabling the system to understand what the wind turbine is experiencing and to distinguish between normal but severe scenarios and abnormal degradation, thus providing a physical background for dynamic benchmarks.

[0017] 3. Dynamic benchmark replaces fixed threshold: The residual is calculated based on the normal prediction value under the current scenario, which greatly reduces the false alarm rate under changing operating conditions and environments.

[0018] 4. Multi-component health index coordination: The independent evaluation of gearbox, generator and converter is integrated with the whole machine, compressing high-dimensional residuals and coupling characteristics into intuitive and unified component health indices, which makes it easy for on-site operation and maintenance personnel to quickly grasp the equipment status and locate it accurately and comprehensively.

[0019] 5. Early Degradation Capture from Multiple Perspectives: Combining trend residuals, multi-source consistency, and environmental triggers, these three indicators complement each other, enabling the early detection of potential anomalies before the health index declines significantly, thus improving the sensitivity of early warning.

[0020] 6. Dynamic correction of remaining lifespan: The degree of coupling risk in the current scenario is used as a correction coefficient to introduce lifespan prediction, realizing the leap from current state assessment to future trend prediction, providing quantitative remaining safe operating time and targeted maintenance decisions, so that the prediction results are more in line with actual operating conditions.

[0021] 7. Closed-loop decision support: The final output includes a complete decision package containing the abnormal location, source of evidence, risk level, remaining lifespan, and maintenance recommendations.

[0022] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a schematic diagram of the method flow provided in an embodiment of the present invention; Figure 2 This is a flowchart of the dynamic identification process for water guiding channels provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of risk warning for water diversion channels provided in an embodiment of the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0027] like Figure 1 As shown, this invention provides a method for dynamic identification and early warning of water-conducting channels in coal mine floor based on microseismic-electric coupling, including: Step 1: Obtain multi-dimensional basic data of the target working face and affected area of ​​the mine, and construct an initial floor water-conducting risk model using the multi-dimensional basic data; Step 1 includes: 1.1 Collect geological structure data, hydrogeological data, mining engineering data, rock physics and geostress data, mining-induced damage prediction data, and existing geophysical and drilling anomaly data of the target working face and its affected area to obtain multidimensional raw data.

[0028] Geological structural data include: the elevation, depth, and undulation of the coal seam floor in the contour map; the precise location, attitude, displacement, width and extension length of faults, fold tracks, and collapse columns exposed in the working face and adjacent areas; and the layer thickness, lithology, sequence stratigraphy, and structural plane development description of each rock layer in the floor in the borehole columnar section.

[0029] Hydrogeological data includes: the stratigraphic position, thickness, top and bottom elevations, water pressure, water level, unit yield, permeability coefficient, storage coefficient, and water quality type of the main aquifers; the recharge, runoff, and discharge conditions and dynamic changes of the aquifers; the hydraulic connection between surface water and the well; the location of historical water inrush points, maximum yield, water source, channel type, and the scope and effectiveness assessment of grouting treatment; and records of water inrush, leakage, and abnormal water temperature in the well boreholes.

[0030] Mining engineering data includes: precise boundaries of the working face, cut-off points, stop lines, and roadway layout plans; designed mining height, mining method, roof management method, expected advance direction, and daily advance rate; and the extent of goaf areas, water accumulation, and mine pressure manifestation data of adjacent working faces.

[0031] Rock physics and geostress data include: uniaxial compressive strength, tensile strength, elastic modulus, Poisson's ratio, cohesion, and internal friction angle of the main rock strata of the base. Measured or inverted geostress magnitudes and directions, particularly the ratio of horizontal to vertical stress.

[0032] Mining-induced failure prediction data includes: estimated values ​​of floor failure depth and their spatial distribution obtained using empirical formulas (such as those based on rock mass strength and mining depth) or numerical simulations.

[0033] Existing geophysical and drilling anomaly data includes: historical geophysical data such as transient electromagnetic, direct current electrical, and audio-frequency electrical imaging delineation of low-resistivity anomaly zones and interpretation results; and borehole verification of the precise location and characteristics of fracture zones, water-bearing anomalies, or cavities revealed.

[0034] 1.2 The multidimensional raw data are uniformly entered into the geological modeling database, and spatial coordinate transformation is performed in the local three-dimensional rectangular coordinate system with the starting point of the target working face as the origin to obtain multidimensional basic data.

[0035] 1.3 Based on the aforementioned multidimensional basic data, 3D geological modeling software is used to construct stratigraphic interface elements, structural entity elements, hydrological attribute elements, and mining space elements to obtain an initial base plate model; specifically including: 1.3.1 Based on the aforementioned multidimensional basic data, using three-dimensional geological modeling software, the coal seam floor interface, the top and bottom interfaces of each aquifer and aquifer are generated sequentially. Fault planes or columns are used to perform three-dimensional characterization of faults and collapse columns, realizing the spatial displacement or absence of strata, and obtaining stratum interface elements. 1.3.2 Faults are interpreted as three-dimensional surface or volumetric elements of attitude, width, and conductivity, and collapse columns are interpreted as columnar anomalies, thus obtaining structural entity elements; 1.3.3 The aquifer water pressure, permeability coefficient and aquitard integrity are taken as continuous attribute fields and interpolated to the entire model space using Kriging or sequential Gaussian simulation to obtain the hydrological attribute volume elements. 1.3.4 Construct a cavity model of the working face cut, two roadways, and planned advance range, and mark adjacent goaf areas to obtain mining space elements; 1.3.5 By combining the stratigraphic interface elements, the structural entity elements, the hydrological attribute elements, and the mining space elements, an initial base plate model is obtained.

[0036] 1.4 Based on the initial base plate model, the study area is discretized into three-dimensional regular hexahedral mesh units, and a three-dimensional index and centroid coordinates are set for each mesh unit.

[0037] 1.5 For each grid cell, quantify the lithology and integrity parameters, structural development parameters, hydraulic distance parameters from the aquifer, equivalent hydrostatic pressure influence coefficient parameters, initial permeability parameters, and mining-induced damage influence parameters to obtain the initial properties of each grid cell and output the initial floor water conduction risk model.

[0038] Step 2: Based on the initial foundation water-conducting risk model, and combining microseismic data and electrodynamic data, design a monitoring network model including monitoring zones and an adaptive monitoring mechanism; Step 2 includes: 2.1 Based on the initial attributes of each grid cell in the initial bottom plate water-conducting risk model, the parameters are weighted and fused to obtain a comprehensive risk score. According to the comprehensive risk score R, the study area is divided into a red key monitoring area, a yellow attention monitoring area, and a green background monitoring area.

[0039] For example: Red key areas R≥0.7: fault fracture zones, collapse columns, areas with high water pressure and thin aquitards, areas around historical water inrush points, and areas within the expected mining damage depth.

[0040] Yellow concern zone 0.4≤R<0.7: Zones with relatively developed fractures, transition zones close to faults or aquifers, and zones with moderate damage to the base plate.

[0041] Green background zone R<0.4: Intact rock strata, a safe area away from structures and water sources.

[0042] 2.2 Based on the divided monitoring areas, microseismic monitoring sensors, electric detection sensors, and hydrological dynamic monitoring sensors with different densities are placed in different areas to obtain an initial sensor deployment scheme. Based on the initial sensor deployment scheme, a forward-moving strategy triggered by mining advancement and a high-density supplementation strategy triggered by anomalies are designed to obtain an adaptive monitoring mechanism and output a monitoring network model.

[0043] The microseismic monitoring sensor includes a downhole three-component microseismic detector, a single-component high-frequency microseismic detector, and an acceleration sensor; the electric detection sensor includes a power supply electrode, a measuring electrode, and a spontaneous potential sensor; and the hydrological dynamic monitoring sensor includes a stratified water pressure sensor, a stratified flow sensor, a hydrological sensor, and an online hydrochemical sensor.

[0044] An example of an initial sensor deployment scheme: 1) Layout of microseismic monitoring Green background zone: On the sidewalls of the roadway outside the working face, a single-component microseismic detector is installed every 80-100 meters to form a sparse skeleton, which is used to capture high-energy signals from a distance and ensure basic positioning capability for the entire mining impact range.

[0045] Yellow Concern Zone: Three-component geophones are installed every 40-60 meters in the tunnels and dedicated monitoring boreholes, and high-frequency geophones are installed in shallow holes drilled 10-30 meters below the floor to improve the detection accuracy of minor fractures in this area.

[0046] Key red zones: Drill deep holes (50-80 meters below the floor) on both sides of the fault and at the edge of the collapse column, and lower multi-stage, three-component geophone strings to achieve three-dimensional surround monitoring. On the floor of the roadways on both sides of the working face in front of key structures, densely distribute shallow-hole geophones every 20-30 meters. Simultaneously deploy microseismic geophones and electrodes in key boreholes to achieve simultaneous measurement and facilitate coupled analysis.

[0047] 2) Electric detection layout Power supply and measurement electrode array: In the green zone, existing anchor bolts or driven electrodes on the tunnel floor and sidewalls are used to construct quadrupole or dipole devices with a large spacing of 30-50 meters, which are periodically excited by low-frequency pulses to perform general survey scans.

[0048] In the yellow area, a ring electrode or wall-mounted electrode array is connected in series inside the borehole in the base plate, with an electrode spacing of 10-20 meters and a measurement electrode spacing of 5 meters, to achieve high-resolution electrical imaging between single holes.

[0049] In the red zone, a combined shaft and tunnel electrode array is deployed around the fault or water-bearing structure: a ring of power supply electrodes is installed in the underground tunnel, and a distributed electrode chain is placed in the boreholes drilling into the floor, such as one electrode every 1-2 meters, forming a three-dimensional electric field focusing body. Multi-frequency sweep excitation polarization observations can be performed in this area to extract polarizability and attenuation spectrum, capturing subtle electrical changes in water-bearing fractures.

[0050] Spontaneous potential and auxiliary sensors: Non-polarizing electrodes are placed in key boreholes to continuously record changes in spontaneous potential, reflecting the filtering potential of groundwater flow. Water temperature probes are integrated at key nodes of the electrode array to help determine the source of water.

[0051] 3) Hydrological monitoring layout Layered hydrological monitoring boreholes were constructed in the red and yellow zones. The aquifer was divided into sections by packers inside the boreholes. Each section was independently equipped with a pressure gauge, a thermometer, and a switchable stratified flow meter to directly obtain changes in water pressure, temperature, and flow rate at different strata.

[0052] Online monitoring devices for total water inflow and water temperature were installed in the goaf and working face drainage ditches.

[0053] The forward movement strategy includes: defining a monitoring window length starting from the opening eye in the working face advancement direction; automatically moving the window forward with each step the working face advances, triggering a sensor switching action; the forward movement action is as follows: The microseismic and electrical resistivity data acquisition host automatically switches channels: it closes the sensor channels behind the window that have entered the goaf, and activates the sensors that have been pre-deployed but not yet activated in front of the window, such as the geophones pre-buried in the borehole further ahead, or the electrode segments that have been buried but are dormant.

[0054] If there are no pre-embedded sensors in front, the system will issue an instruction to remind users to quickly install and connect a new round of short-hole sensors within a specified distance in front of the window, ensuring that there is always a high-density monitoring area covering the base plate 60-100 meters in front of the work area.

[0055] In this way, the center of gravity of the monitoring network shifts synchronously with the mining-affected area, ensuring that the areas with the most severe stress disturbances and the most active ruptures are always covered with high density.

[0056] The high-density replenishment strategy includes: by setting an anomaly threshold, when the energy release rate or frequency of microseismic events in a certain area suddenly increases, or the apparent resistivity decreases by more than 10%, or the spontaneous potential shows a sustained drift of more than 5mV, the area is automatically designated as a risk activation zone, and the backup sensor activation action is triggered. The backup sensor activation action is as follows: The system selects available tunnels or boreholes around the risk activation zone and uses an automatic switching matrix, such as a multi-electrode switcher, to incorporate nearby backup electrodes or sensors into the active acquisition network, thereby achieving virtual encryption.

[0057] If the physical location cannot be covered by switching, the system issues an instruction to quickly drill 1-2 shallow holes in the area, insert short-section detectors and electrode rods, and quickly connect to the monitoring network wirelessly or via wired means for enhanced monitoring for 48 hours.

[0058] During the enhancement period, the trigger sensitivity of microseismic acquisition in the area was increased, and the electrical observation mode was switched to continuous high-frequency multi-frequency excitation to obtain denser time sampling.

[0059] Step 3: Collect microseismic spatiotemporal response data using the monitoring network model to obtain the base plate crack evolution feature set; Step 3 includes: 3.1 Continuous data is collected using the monitoring network model. The collected continuous data is automatically denoised, and waveform segments of suspected microseismic events are extracted. Longitudinal and transverse waves are extracted from the waveform segments of suspected microseismic events to obtain sub-millisecond microseismic events.

[0060] 3.2 Based on the microseismic event, the location and time of rupture occurrence are inverted using the initial bottom plate water-conducting risk model to obtain the rupture source of the microseismic event. Then, for the microseismic event, the microseismic energy, dominant frequency and spectral characteristics, moment tensor inversion and rupture type, and main direction of crack propagation are calculated to obtain the microseismic event file.

[0061] Microseismic energy: The relative energy released during a rupture is calculated by integrating the amplitude and duration of the waveform, usually expressed in Joules or their logarithmic form. Higher energy indicates a larger or wider rupture.

[0062] Dominant frequency and spectral characteristics: Perform a Fast Fourier Transform on the waveform to identify the frequency with the highest energy concentration in the signal. Generally, tensional fractures involve large volume changes and often excite signals rich in low-frequency components; while small-scale shear slips tend to have higher dominant frequencies. The width and peak value variations of the spectrum can reflect the scale of the fracture and whether fluid is involved.

[0063] Moment tensor inversion and fracture types: The moment tensor describes the distribution of equivalent forces at a seismic source at an instant. Using the initial direction and amplitude ratio of the P-waves recorded by multiple sensors, the moment tensor matrix is ​​obtained through inversion. After decomposition, the proportions of shear, tensile, and explosive / inward bursting components can be calculated. This is analogous to analyzing ground vibration patterns to deduce whether the subsurface was torn apart or displaced.

[0064] Principal direction of fracture propagation: Based on the moment tensor solution, two possible orientations (strike and dip) of the fracture surface can be determined. Combined with the orientation of regional geological structures or the direction of the stress field, the most likely direction of fracture propagation can be inferred. Furthermore, for a group of events located close together, principal component analysis (PCA) of spatial point clouds can also be used to calculate the direction of their major axis, thus obtaining the principal direction of macroscopic fracture propagation.

[0065] 3.3 Based on the microseismic events, a moving time window is defined. Within the moving time window, it is identified whether there is a sudden increase in event frequency, a continuous increase in event energy, a movement of the average event depth towards deeper aquifers, an directional arrangement of events along known faults in the initial bottom plate water-conducting risk model, a sudden increase in the proportion of tensional events, a significant increase in the proportion of low-frequency events, and a continuous expansion of the event spatial volume, thereby obtaining abnormal signals.

[0066] A sudden increase in event frequency: If the number of events in this window suddenly increases by two or three times compared to the average of the previous few windows, it indicates that the rock mass is fracturing at an accelerated rate, which is a dangerous precursor.

[0067] Event energy continues to increase: The cumulative value of all event energy within the window is getting larger and larger, especially the appearance of high-energy events, indicating that small cracks are connecting to form a large fracture.

[0068] The average depth of events is shifting towards deeper aquifers: The average depth or deepest event depth within a window is compared to the previous window. If the queue of rupture points is observed moving step by step from the shallow bottom towards deeper aquifers, this is the most direct alarm.

[0069] Events are oriented along known faults in the initial base plate water-conducting risk model: if the events are projected onto a 3D graph and they suddenly align in a line or a narrow band along known fault or collapse column boundaries in the model, it indicates that the old structure has been activated and has become a potential water-conducting channel skeleton.

[0070] A sudden increase in the proportion of tensional fracturing events: Under normal circumstances, rock fracture under pressure is primarily caused by shear. However, within a certain window, if we observe a sudden increase in the proportion of tensional fracturing events, this usually means that the fracture is no longer simply shifting, but rather being stretched. The reason for this stretching is likely that water has entered the fracture, and the water pressure has forced the crack open. This is a very critical fluid-structure interaction signal.

[0071] A significant increase in the proportion of low-frequency events: Since the flow and oscillation of fluid in the fracture will generate unique low-frequency signals, if the proportion of low-frequency events in the window increases significantly, even if it does not directly prove the presence of water, it strongly suggests that there is saturated fluid in the fracture.

[0072] The event space volume continues to expand: Calculate the volume of the smallest three-dimensional convex hull or ellipsoid that can enclose all events. If the volume continues to expand rapidly like an inflating balloon, it indicates that the fracture network is spreading over a large area, increasing the risk of connecting multiple water sources.

[0073] 3.4 Based on the anomalous signal, risk quantification is performed to calculate the crack activity score and obtain the base plate crack evolution index. Then, the microseismic time archive and the base plate crack evolution index are integrated to obtain the base plate crack evolution feature set.

[0074] Step 4: Based on the base plate fracture evolution feature set, trigger zoned electric excitation and response acquisition to obtain the water-bearing fracture electric response feature set; Step 4 includes: 4.1 For the base plate crack evolution index, a concern threshold is set to determine whether the base plate crack evolution index of each grid cell exceeds the concern threshold. If so, it is determined that shear fracture, tension fracture or crack is extending into the aquifer and triggers the physical examination threshold. The area to be inspected is delineated by expanding outward from the grid cell as the center.

[0075] 4.2 Based on the area to be inspected, the base plate crack evolution index is used to perform low-frequency pulse electric field excitation, multi-frequency sweep excitation, excitation polarization attenuation measurement, natural potential dynamic monitoring, and zoned cross power supply measurement to complete multi-mode joint electric excitation.

[0076] Low-frequency pulsed electric field excitation: A brief low-frequency high-voltage pulse is instantaneously released onto a set of power supply electrodes, and then the voltage decay curve is recorded using surrounding measuring electrodes.

[0077] Multi-frequency sweep excitation: Instead of emitting current at only one frequency, it continuously emits a series of alternating currents from high frequency to low frequency.

[0078] Excitation polarization decay measurement: First, apply DC power for a while, then suddenly disconnect the power supply and observe how the voltage slowly disappears.

[0079] Dynamic monitoring of natural potential: without any power supply, the weak electric field naturally existing in the earth is measured using only non-polarized electrodes.

[0080] Zoned cross-power supply measurement: Switch different power supply electrode pairs to allow current to pass through the suspected area from different directions.

[0081] 4.3 Based on the data collected by the multi-mode joint electric excitation, inversion is performed to calculate the real-time resistivity, polarizability, excitation polarization decay time constant, pulse electric field response amplitude, natural potential change, and electric response phase difference, thereby obtaining the real-time electric response parameters.

[0082] 4.4 The real-time electrodynamic response parameters are compared with the original electrical background values ​​in the initial bottom plate water-conducting risk model to calculate the resistivity change and the spontaneous potential change. The changes are weighted and fused to obtain the hydrophobic fissure electrodynamic activation index. Based on the hydrophobic fissure electrodynamic activation index, the hydrophobic fissure electrodynamic response feature set of the study area is output. The hydrophobic fissure electrodynamic response feature set includes resistivity change, excitation polarization characteristics of hydrophobic fissures, spontaneous potential orientation change, and impulse response.

[0083] Step 5, as follows Figure 2 As shown, using the base plate fracture evolution feature set and the water-bearing fracture electrodynamic response feature set, a spatiotemporal coupled field combining microseismic and electrodynamic data is constructed to obtain a candidate water-conducting channel set; step 5 includes: 5.1 The evolution feature set of the base plate cracks and the electrodynamic response feature set of the water-bearing cracks are mapped to the initial base plate water-conducting risk model to unify the spatiotemporal coordinate system. In each grid cell, the microseismic data and electrodynamic data are compared to calculate the spatial overlap of the anomalies and the time difference of the anomaly occurrence.

[0084] 5.2 Determine whether the spatial overlap and time difference of the grid cells meet the preset range standard. If so, according to the main direction of fracture propagation, the direction of electrodynamic anomaly gradient and the direction pointing to the aquifer, the grid cells that meet the connectivity, directional consistency and temporal logic are connected into an ordered set by the shortest path algorithm to obtain the candidate water-conducting channel.

[0085] 5.3 Calculate the starting and ending points, direction and tendency, extension length, volume and dynamic evolution rate of the candidate water-conducting channels for each day to construct a candidate water-conducting channel profile.

[0086] Step 6: Using the candidate water-diverting channels and hydrological dynamic data, determine whether the water-diverting channels are connected. If so, calculate the water-diverting capacity to obtain a channel water-diverting capacity evaluation set. Step 6 includes: 6.1 For each of the candidate water diversion channels, obtain the water source and outlet, and extract all hydrological sensor data in the area where the candidate water diversion channel is located to calculate water pressure changes, stratified flow changes, and water temperature and water chemistry tracers to obtain hydrological dynamic data.

[0087] 6.2 Based on the water source, outlet, and hydrological dynamic data, the probability of the candidate water diversion channel's connection is calculated using the Sigmoid function. The channel's connection status is then determined according to the magnitude of the probability. The channel status includes: Not connected: There are cracks and water in some areas, but they are not connected and are not forming a line. Potential connection: If there are intermittent but slight connections, or if there are minor hydrological changes, vigilance and close monitoring are needed. Critical breakthrough: The channel is basically formed, the water pressure is dropping, or the flow rate is increasing, and the actual water inrush may only be one violent mine pressure disturbance away; Completed: An effective hydraulic connection from the aquifer to the mining space has been or is nearing completion.

[0088] 6.3 Based on the channel's connectivity status, an equivalent water-conducting model of the candidate water-conducting channel is constructed using Darcy's law. The equivalent permeability coefficient is inverted using soft computing, the equivalent water-passing area is calculated, the water pressure difference and length are obtained, and the water-conducting direction and potential water inrush location are determined to form a water-conducting capacity evaluation set for the water-conducting channel. The water-conducting capacity evaluation set includes the channel ID and spatial location, geometric dimensions, fracture development degree, water filling degree, connectivity probability, equivalent water-conducting capacity, potential water inrush location, and risk development rate.

[0089] Step 7: Utilize the aforementioned water-conducting capacity evaluation set to classify risk levels, and implement corresponding early warnings and differentiated management for different risk levels. Step 7 includes: 7.1 Based on the water-conducting capacity evaluation set, parameter weighting and fusion are performed to obtain the risk index, as shown in 3. The risk level is divided according to the risk index to obtain a blue warning indicating that the bottom plate is generally calm, a yellow warning indicating that there are signs of activity, an orange warning indicating that there is a clear danger and a risk of deterioration, and a red warning indicating that there is a high probability of water inrush.

[0090] 7.2 When the risk level is determined to be a blue alert, no risk measures shall be taken.

[0091] 7.3 When the risk level is determined to be a yellow alert, increase the frequency of microseismic and electrodynamic data acquisition, send the specific locations that need to be verified by intensive drilling, and remind the team to execute within a time limit.

[0092] 7.4 When the risk level is determined to be an orange alert, a speed limit command is sent directly to the coal mining machine and the dispatch center, or it is recommended to partially stop mining and activate the preset water exploration or grouting plan.

[0093] 7.5 When the risk level is determined to be a red alert, the underground audible and visual alarm will be triggered immediately, the non-essential power supply to the dangerous area will be automatically cut off, the evacuation route map will be pushed to all relevant personnel, and emergency response measures will be initiated simultaneously through the ground command center.

[0094] Therefore, this invention provides a dynamic identification and early warning method for water diversion channels in coal mine floor based on microseismic-electric coupling. By extracting coupling features of the nacelle environment and operating conditions and identifying the scene, a dynamic health benchmark and a multi-dimensional residual analysis system are constructed to achieve integrated intelligent monitoring of the health status assessment, early degradation warning and dynamic prediction of remaining life of key components of wind turbines.

[0095] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0096] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for dynamic identification and early warning of water diversion channels in coal mine floor based on microseismic-electric coupling, characterized in that, include: Acquire multidimensional basic data of the target working face and the affected area of ​​the mine, and use the multidimensional basic data to construct an initial floor water-conducting risk model; Based on the initial bottom plate water-conducting risk model, and combined with microseismic data and electrodynamic data, a monitoring network model including monitoring zones and an adaptive monitoring mechanism is designed. The monitoring network model is used to collect microseismic spatiotemporal response data to obtain a set of characteristics of the evolution of base plate cracks; Based on the aforementioned base plate fracture evolution feature set, regional electric excitation and response acquisition are triggered to obtain the water-bearing fracture electric response feature set; Using the base plate crack evolution feature set and the water-bearing crack electrodynamic response feature set, a spatiotemporal coupled field combining microseismic and electrodynamic data is constructed to obtain a candidate water-conducting channel set; Using the candidate water diversion channels and hydrological dynamic data, it is determined whether the water diversion channels are connected. If so, the water diversion capacity is calculated to obtain the channel water diversion capacity evaluation set. The risk level is classified using the water conduction capacity evaluation set, and corresponding early warning and differentiated management are carried out for different risk levels.

2. The method for dynamic identification and early warning of water diversion channels in coal mine floor based on microseismic-electric coupling according to claim 1, characterized in that, Acquire multidimensional basic data of the target working face and its affected area in the mine, and use the multidimensional basic data to construct an initial floor water-conducting risk model, including: Collect geological structure data, hydrogeological data, mining engineering data, rock physics and geostress data, mining-induced damage prediction data, and existing geophysical and drilling anomaly data of the target working face and its affected area to obtain multidimensional raw data; The multidimensional raw data are uniformly entered into the geological modeling database, and spatial coordinate transformation is performed in the local three-dimensional rectangular coordinate system with the starting point of the target working face as the origin to obtain multidimensional basic data. Based on the aforementioned multidimensional basic data, 3D geological modeling software is used to construct stratigraphic interface elements, structural entity elements, hydrological attribute elements, and mining space elements to obtain an initial base plate model. Based on the initial base plate model, the study area is discretized into three-dimensional regular hexahedral grid cells, and a three-dimensional index and centroid coordinates are set for each grid cell; For each grid cell, quantify the lithology and integrity parameters, structural development parameters, hydraulic distance parameters from the aquifer, equivalent hydrostatic pressure influence coefficient parameters, initial permeability parameters, and mining-induced damage influence parameters to obtain the initial properties of each grid cell and output the initial floor water conduction risk model.

3. The method for dynamic identification and early warning of coal mine floor water diversion channels based on microseismic-electric coupling according to claim 2, characterized in that, Based on the aforementioned multidimensional basic data, 3D geological modeling software is used to construct stratigraphic interface elements, structural entity elements, hydrological attribute elements, and mining space elements to obtain an initial base model, including: Based on the aforementioned multidimensional basic data, using three-dimensional geological modeling software, the coal seam floor interface, the top and bottom interfaces of each aquifer and aquifer are generated sequentially, and the fault plane or column is used to perform three-dimensional characterization of the fault and collapse column to obtain the stratigraphic interface elements. Faults are interpreted as three-dimensional surface or volumetric features of attitude, width, and conductivity, and collapse columns are interpreted as columnar anomalies, thus obtaining structural entity features. The aquifer water pressure, permeability coefficient and aquitard integrity are treated as continuous attribute fields and interpolated to the entire model space using Kriging or sequential Gaussian simulations to obtain the hydrological attribute volume elements. Construct a cavity model of the working face cut, two roadways, and planned advance range, and mark adjacent goaf areas to obtain mining space elements; By combining the stratigraphic interface elements, the structural entity elements, the hydrological attribute elements, and the mining space elements, an initial base plate model is obtained.

4. The method for dynamic identification and early warning of water diversion channels in coal mine floor based on microseismic-electric coupling according to claim 3, characterized in that, Based on the initial foundation water-conducting risk model, and combining microseismic data and electrodynamic data, a monitoring network model including monitoring zones and an adaptive monitoring mechanism is designed, comprising: Based on the initial attributes of each grid cell in the initial bottom plate water-conducting risk model, the parameters are weighted and fused to obtain a comprehensive risk score. According to the comprehensive risk score, the study area is divided into a red key monitoring area, a yellow attention monitoring area, and a green background monitoring area. Based on the divided monitoring areas, microseismic monitoring sensors, electric detection sensors, and hydrological dynamic monitoring sensors with different densities are placed in different areas to obtain an initial sensor deployment scheme. Based on the initial sensor deployment scheme, a forward-moving strategy triggered by mining advancement and a high-density supplementation strategy triggered by anomalies are designed to obtain an adaptive monitoring mechanism and output a monitoring network model. The microseismic monitoring sensor includes a downhole three-component microseismic detector, a single-component high-frequency microseismic detector, and an acceleration sensor; the electric detection sensor includes a power supply electrode, a measuring electrode, and a spontaneous potential sensor; and the hydrological dynamic monitoring sensor includes a stratified water pressure sensor, a stratified flow sensor, a hydrological sensor, and an online hydrochemical sensor.

5. The method for dynamic identification and early warning of water diversion channels in coal mine floor based on microseismic-electric coupling according to claim 4, characterized in that: The forward movement strategy includes: defining a monitoring window length starting from the opening eye in the working face advancement direction; the window automatically moves forward with each step the working face advances, and triggers a sensor switching action. The high-density replenishment strategy includes: by setting an abnormal threshold, when the energy release rate or frequency of microseismic events in a certain area suddenly increases, or the apparent resistivity decreases by more than 10%, or the natural potential shows a continuous drift of more than 5mV, the area is automatically labeled as a risk activation zone and the backup sensor activation action is triggered.

6. The method for dynamic identification and early warning of coal mine floor water guiding channels based on microseismic-electric coupling according to claim 5, characterized in that, By using the aforementioned monitoring network model to collect microseismic spatiotemporal response data, a set of characteristics for the evolution of foundation cracks is obtained, including: The monitoring network model is used to collect continuous data. The collected continuous data is automatically denoised, and waveform segments of suspected microseismic events are extracted. Longitudinal and transverse waves are extracted from the waveform segments of suspected microseismic events to obtain sub-millisecond microseismic events. Based on the microseismic events, the location and time of rupture occurrence are inverted using the initial bottom plate water-conducting risk model to obtain the rupture source of the microseismic events. Then, for the microseismic events, the microseismic energy, dominant frequency and spectral characteristics, moment tensor inversion and rupture type, and main direction of crack propagation are calculated to obtain the microseismic event archive. Based on the microseismic events, a moving time window is defined. Within the moving time window, it is identified whether there is a sudden increase in event frequency, a continuous increase in event energy, a movement of the average event depth towards the deeper aquifer, an directional arrangement of events along known faults in the initial bottom plate water-conducting risk model, a sudden increase in the proportion of tensional events, a significant increase in the proportion of low-frequency events, and a continuous expansion of the event spatial volume, thereby obtaining abnormal signals. Based on the anomalous signals, risk quantification is performed to calculate the crack activity score and obtain the base plate crack evolution index. Then, the microseismic time archive and the base plate crack evolution index are integrated to obtain the base plate crack evolution feature set.

7. A method for dynamic identification and early warning of coal mine floor water diversion channels based on microseismic-electric coupling according to claim 6, characterized in that, Based on the aforementioned base plate fracture evolution feature set, zonal electric excitation and response acquisition are triggered to obtain the water-bearing fracture electric response feature set, including: For the aforementioned base plate crack evolution index, a focus threshold is set to determine whether the base plate crack evolution index of each grid cell exceeds the focus threshold. If so, it is determined that shear fracture, tension fracture, or crack is extending into the aquifer and triggering the physical examination threshold. The area to be examined is then delineated by expanding outward from the grid cell as the center. Based on the area to be inspected, the base plate crack evolution index is used to perform low-frequency pulse electric field excitation, multi-frequency sweep frequency excitation, excitation polarization attenuation measurement, natural potential dynamic monitoring and zoned cross power supply measurement to complete multi-mode joint electric excitation; Based on the data collected by the multi-mode joint electric excitation, inversion is performed to calculate the real-time resistivity, polarizability, excitation polarization decay time constant, pulse electric field response amplitude, natural potential change, and electric response phase difference, thereby obtaining the real-time electric response parameters. The real-time electrodynamic response parameters are compared with the original electrical background values ​​in the initial bottom plate water-conducting risk model to calculate the resistivity change and spontaneous potential change. The changes are weighted and fused to obtain the hydrodynamic activation index of the aquifer. Based on the hydrodynamic activation index, the hydrodynamic response feature set of the aquifer in the study area is output. The hydrodynamic response feature set of the aquifer includes resistivity change, excitation polarization characteristics of the aquifer, directional change of spontaneous potential, and impulse response.

8. A method for dynamic identification and early warning of water diversion channels in coal mine floor based on microseismic-electric coupling according to claim 7, characterized in that, Using the base plate fracture evolution feature set and the water-bearing fracture electrodynamic response feature set, a spatiotemporal coupled field combining microseismic and electrodynamic data is constructed to obtain a candidate water-conducting channel set, including: The evolution feature set of the base plate cracks and the electrodynamic response feature set of the water-bearing cracks are mapped to the initial base plate water-conducting risk model to unify the spatiotemporal coordinate system. In each grid cell, the microseismic data and electrodynamic data are compared to calculate the spatial overlap of the anomalies and the time difference of the anomaly occurrence. Determine whether the spatial overlap and time difference of the grid cells meet the preset range standard. If so, according to the main direction of fracture propagation, the direction of electrodynamic anomaly gradient and the direction pointing to the aquifer, the grid cells that meet the connectivity, directional consistency and temporal logic are connected into an ordered set by the shortest path algorithm to obtain the candidate water-conducting channel. The starting and ending points, direction and tendency, extension length, volume and dynamic evolution rate of the candidate water-conducting channels are calculated daily to construct a candidate water-conducting channel profile.

9. A method for dynamic identification and early warning of water diversion channels in coal mine floor based on microseismic-electric coupling according to claim 8, characterized in that, Using the candidate water-diverting channels and hydrological dynamic data, it is determined whether the water-diverting channels are connected. If so, the water-diverting capacity is calculated to obtain a channel water-diverting capacity evaluation set, including: For each candidate water diversion channel, the water source and outlet are acquired, and all hydrological sensor data in the area where the candidate water diversion channel is located are extracted to calculate water pressure changes, stratified flow changes, and water temperature and water chemistry tracers to obtain hydrological dynamic data. Based on the water source, outlet, and hydrological dynamic data, the probability of the candidate water diversion channel being connected is calculated using the Sigmoid function. According to the magnitude of the connection probability, the channel connection status is divided into: not connected, potentially connected, critically connected, and connected. Based on the channel's connectivity status, an equivalent water-conducting model of the candidate water-conducting channel is constructed using Darcy's law. The equivalent permeability coefficient is then inverted using soft computing, the equivalent water-passing area is calculated, the water pressure difference and length are obtained, and the water-conducting direction and potential water inrush location are determined to form a water-conducting capacity evaluation set for the water-conducting channel. This evaluation set includes the channel ID and spatial location, geometric dimensions, fracture development degree, water filling degree, connectivity probability, equivalent water-conducting capacity, potential water inrush location, and risk development rate.

10. A method for dynamic identification and early warning of water diversion channels in coal mine floor based on microseismic-electric coupling according to claim 9, characterized in that, The aforementioned water-conducting capacity evaluation set is used to classify risk levels, and corresponding early warnings and differentiated management are implemented for different risk levels, including: Based on the water-conducting capacity evaluation set, parameter weighting and fusion are performed to obtain a risk index. Based on the risk index, the risk level is divided into blue warnings (indicating overall calm bottom), yellow warnings (indicating signs of activity), orange warnings (indicating clear danger and risk of deterioration), and red warnings (indicating a high probability of water inrush). When the risk level is determined to be a blue alert, no risk measures will be taken; When the risk level is determined to be a yellow alert, the frequency of microseismic and electrodynamic data acquisition is increased, and the specific locations that need to be verified by intensive drilling are sent to remind the team to execute within a time limit. When the risk level is determined to be an orange alert, a speed limit command is sent directly to the coal mining machine and the dispatch center, or it is recommended to partially stop mining and activate the preset water exploration or grouting plan. When the risk level is determined to be a red alert, the underground audible and visual alarm will be triggered immediately, the unnecessary power supply to the dangerous area will be automatically cut off, the evacuation route map will be pushed to all relevant personnel, and emergency response measures will be initiated simultaneously through the ground command center.

Images