Water disaster detection method for preventing and controlling water on coal mine working face

Abstract

The invention relates to a water disaster detection method for preventing and controlling water on a coal mine working face. The method comprises the following steps: constructing an initial geologic model containing three-dimensional distribution of an aquifer by acquiring a geologic structure, lithologic characteristics and drilling data; identifying the spatial positions of the fault and the aquifer through a geophysical prospecting method and acquiring a permeability coefficient; integrating the water level dynamic monitoring data to determine goaf water distribution characteristics; constructing a prediction model based on the initial model in combination with the permeability coefficient and goaf water parameters, and outputting a multi-category hidden danger probability distribution diagram; a water disaster risk high-incidence area identification result is generated through spatial correlation analysis; and classifying water disaster types and hazard grades by utilizing clustering. The method has the technical effects that a prediction model capable of comprehensively reflecting water disaster hidden danger distribution is constructed to be used for a water disaster detection link of coal mine working face water prevention and control, and multi-dimensional accurate identification and dynamic risk grading of the water disaster hidden danger of the coal mine working face are achieved.

Description

Technical Field

[0001] The invention belongs to the field of mining engineering and safety technology, and in particular relates to a water hazard detection method for preventing and controlling water in a coal mine working face. Background Art

[0002] A key technical challenge in water hazard detection for coal mine working faces is how to accurately predict the spatial distribution of different types of water hazards within the working face under complex geological conditions. The working face has complex geological structures and often contains multiple types of water hazards, including old goaf water, fault water, and aquifer water. The distribution of these water hazards is closely related to the geological structure and hydrogeological conditions.

[0003] However, existing exploration methods, such as drilling, geophysical exploration, and monitoring data, each have their limitations. Drilling can obtain localized, accurate hydrogeological information, but it is costly and has limited coverage. Geophysical exploration can survey large areas, but its resolution is low, making it difficult to accurately identify small structures or localized aquifers. Monitoring data, while able to reflect hydrological dynamics in real time, lacks information on spatial distribution.

[0004] Therefore, how to comprehensively use these data to construct a prediction model that can comprehensively reflect the distribution of water hazard risks for use in the water hazard detection link of coal mine working face water prevention and control has become a technical difficulty. Summary of the Invention

[0005] Based on this, it is necessary to provide a water hazard detection method for preventing and controlling water in coal mine working faces in response to the above technical problems.

[0006] In a first aspect, the present application provides a method for detecting water hazards in a coal mine working face for preventing and controlling water, comprising:

[0007] S1: Construct an initial geological model based on the working face structural data, lithologic characteristics data, and drilling data. The initial geological model contains the three-dimensional spatial distribution information of the aquifer.

[0008] S2: Based on the working face structural data, geophysical exploration methods are used to obtain the spatial position data of the fault and the aquifer, and the permeability coefficient of the fault water and the permeability coefficient of the aquifer water;

[0009] S3: Obtain the characteristic parameters of old empty water based on the water level dynamic monitoring data and regional hydrogeological conditions data; the characteristic parameters of old empty water include distribution range, supply source and water quality characteristics;

[0010] S4: Based on the initial geological model, combined with the spatial location data of faults, the spatial location data of aquifers, the fault water permeability coefficient, the aquifer water permeability coefficient, and the characteristic parameters of old water, a water hazard distribution prediction model is constructed; the water hazard distribution prediction model is used to classify and predict water hazard hazards, and a multi-category hazard probability distribution map is obtained;

[0011] S5: Perform spatial correlation analysis on the probability distribution map of multi-category hidden dangers, the spatial location data of faults, and the spatial location data of aquifers to generate identification results of high-risk areas for flood hazards;

[0012] S6: Based on the identification results, cluster the areas with high flood risk to obtain clustering results; based on the clustering results, determine the types and severity of flood damage in different areas.

[0013] In a second aspect, the present application also provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements a water hazard detection method for preventing and controlling water in a coal mine working face as described in the first aspect.

[0014] In a third aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a water hazard detection method for preventing and controlling water in a coal mine working face as described in the first aspect.

[0015] The above-mentioned method for detecting water hazards in coal mine working faces involves constructing an initial geological model containing the three-dimensional distribution of aquifers by acquiring geological structures, lithologic characteristics, and drilling data; identifying the spatial locations of faults and aquifers and obtaining permeability coefficients through geophysical methods; integrating water level dynamic monitoring data to determine the distribution characteristics of old empty water; constructing a prediction model based on the initial model combined with permeability coefficients and old empty water parameters, and outputting a multi-category hidden danger probability distribution map; generating identification results for high-risk areas for water hazards through spatial correlation analysis; and using clustering to classify water hazard types and hazard levels. Its technical effect is to construct a prediction model that can comprehensively reflect the distribution of water hazard hazards for use in the water hazard detection phase of coal mine working face water prevention and control, thereby achieving multi-dimensional and accurate identification of water hazard hazards in coal mine working faces and dynamic risk grading. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 A schematic flow chart of a method for detecting water hazards in a coal mine working face for preventing and controlling water provided by the present invention;

[0018] Figure 2 The figure is a flowchart of the steps for generating a multi-category hidden danger probability distribution map in an optional embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0020] refer to Figure 1 , which presents a flow chart of a water hazard detection method for preventing and controlling water in a coal mine working face provided by the present application, the method comprising the following steps:

[0021] S1: Based on the working face structural data, lithologic characteristic data, and drilling data, an initial geological model is constructed. The initial geological model contains the three-dimensional spatial distribution information of the aquifer.

[0022] Specifically, structural data for the working face can be collected from multiple sources. This includes information on structural features such as folds and faults from geological exploration reports, including specific data such as the axial direction and dip of folds and the strike and drop of faults. Lithologic data relates to the types of different rock formations within the working face area, such as sandstone, mudstone, and limestone, as well as their physical properties, such as hardness, porosity, and permeability. Drilling data is crucial, providing stratigraphic information acquired during the actual drilling process, including borehole depth, diameter variations, and core sampling.

[0023] Professional geological modeling software can be used to digitize the collected data and construct the three-dimensional distribution of the aquifer. This distribution includes not only the aquifer's planar position but also its vertical thickness variations, the undulations of the roof and floor, and so on. For example, through mathematical methods such as interpolation algorithms, the spatial distribution of the aquifer within the entire working area can be inferred based on the depth data of the aquifer in the borehole, allowing geologists to intuitively visualize the three-dimensional shape of the aquifer in the underground space.

[0024] S2: Based on the working face structural data, geophysical exploration methods are used to obtain the fault spatial position data and the aquifer spatial position data, and the fault water permeability coefficient and the aquifer water permeability coefficient are obtained.

[0025] Specifically, appropriate geophysical methods can be selected based on the specific geological conditions of the working face and the purpose of the exploration. Geophysical methods can include seismic exploration, electromagnetic exploration, and the like. Seismic exploration involves artificially exciting seismic waves, then receiving and analyzing the characteristics of the seismic waves propagating in the underground medium, such as changes in wave velocity, amplitude, and phase, to infer underground geological structures and rock formation properties. Electromagnetic exploration utilizes the propagation and distribution patterns of electromagnetic fields in the underground medium, and measures parameters such as the intensity and frequency of the electromagnetic field to detect the electrical characteristics of underground geological bodies, thereby indirectly obtaining geological information. During implementation, appropriate detection equipment, such as seismic detectors, electromagnetic transmitting and receiving devices, and the like, can be deployed to collect data in the working face area at a certain grid spacing.

[0026] After acquiring a large amount of raw data through geophysical exploration methods, specialized data processing software and algorithms are used to process and analyze this data, extracting the spatial location information of faults and aquifers. Furthermore, combining relevant geological theories and empirical formulas, the fault and aquifer water permeability coefficients are calculated. These permeability coefficients reflect the underground medium's ability to conduct water flow and can be used to assess water hazard risks. For example, by analyzing changes in the propagation characteristics of seismic waves near a fault, the fault's location and direction can be determined. Combined with laboratory rock permeability data and actual field conditions, the fault's water permeability coefficient can be calculated.

[0027] S3: Based on the water level dynamic monitoring data and regional hydrogeological conditions data, the characteristic parameters of the old empty water are obtained; the characteristic parameters of the old empty water include distribution range, supply source and water quality characteristics.

[0028] Specifically, water level monitoring points can be set up around the working face and in areas potentially affected by old water to monitor groundwater level changes regularly or in real time. These monitoring points can use automatic water level monitoring equipment to accurately record dynamic water level changes, including the magnitude and rate of change. Long-term monitoring and analysis of this data can help us understand the dynamic patterns of groundwater changes.

[0029] Collect data on the hydrogeological conditions of the working face area, including the sources of groundwater recharge in the area, such as atmospheric precipitation and the recharge of surface water bodies; groundwater runoff conditions, such as groundwater flow direction and flow rate; and groundwater discharge methods, such as discharge through springs, rivers and other surface water bodies. At the same time, obtain hydrogeological parameters in the area, such as groundwater storage capacity, water supply degree, and overflow coefficient. Through comprehensive analysis of the above data, combined with dynamic water level monitoring data, the characteristic parameters of the old empty water can be accurately determined, including its distribution range, recharge source, and water quality characteristics. For example, based on the precipitation data and groundwater runoff direction in the area, the recharge source and possible distribution area of ​​the old empty water can be inferred. Combined with water quality analysis data, the water quality of the old empty water, such as pH and mineralization, can be understood.

[0030] S4: Based on the initial geological model, combined with the fault spatial location data, aquifer spatial location data, fault water permeability coefficient, aquifer water permeability coefficient and old empty water characteristic parameters, a water hazard distribution prediction model is constructed; the water hazard distribution prediction model is used to classify and predict water hazard hazards, and a multi-category hazard probability distribution map is obtained.

[0031] Specifically, using the initial geological model constructed in step S1 as the basic framework, the spatial position data of the faults and aquifers, the fault water permeability coefficients and aquifer water permeability coefficients obtained in step S2, and the characteristic parameters of the old water gap obtained in step S3 are used as input data and key parameters. Using computer simulation technology and numerical calculation methods, a water hazard distribution prediction model is constructed. This model comprehensively considers multiple factors, including geological structure, lithologic characteristics, groundwater permeability characteristics, and the presence of old water gaps. It can simulate the flow and distribution of groundwater within the working face area, thereby predicting areas where water hazards may occur.

[0032] By running a water hazard distribution prediction model, a classification forecast of water hazards within the working face area is conducted. Based on the prediction results, water hazard hazards are divided into different categories, such as high-risk hazards, medium-risk hazards, and low-risk hazards. Visualization technologies such as geographic information systems (GIS) can then be used to present these hazard categories in the form of probability distribution maps. Different colors or symbols can be used on the hazard probability distribution map to represent hazard areas of different risk levels, allowing geologists and managers to intuitively understand the spatial distribution of water hazard hazards within the working face area.

[0033] S5: Perform spatial correlation analysis on the probability distribution map of multi-category hidden dangers, fault spatial location data, and aquifer spatial location data to generate identification results of high-risk areas for water disasters.

[0034] Specifically, the multi-category hazard probability distribution map generated in step S4 is overlaid with the fault and aquifer spatial location data from step S2, and spatial correlation analysis is performed. Professional GIS software or spatial analysis tools can be used to analyze the spatial relationship between hazard areas, faults, and aquifers using spatial query and spatial statistics. For example, the distance between hazard areas and faults, the intersection of faults and aquifers, and the distribution of hazard areas within aquifers can be calculated, thereby deriving spatial correlation characteristics between water hazards, geological structures, and aquifers.

[0035] Based on the results of spatial correlation analysis, high-risk areas for flooding are identified. These areas are characterized by a high probability of hidden dangers and are closely related to the spatial location of faults and aquifers. Examples include areas near the intersection of faults and aquifers, or areas within aquifers close to faults. By identifying these high-risk areas, we can more accurately identify key areas for flood control and provide targeted guidance for the development of subsequent flood control measures.

[0036] S6: Based on the identification results, cluster the areas with high flood risk to obtain clustering results; based on the clustering results, determine the types and severity of flood damage in different areas.

[0037] Specifically, based on the identification results of high-risk flood areas obtained in step S5, a cluster analysis algorithm is used to cluster these high-risk areas. Cluster analysis is an unsupervised machine learning method that automatically divides data into different categories based on similarities and differences. In this step, cluster analysis can be performed based on multiple dimensions of data, such as the probability of hidden dangers in high-risk areas, the spatial relationship with faults and aquifers, and the characteristic parameters of old water. High-risk areas with similar characteristics can be grouped into the same category.

[0038] By analyzing the clustering results and combining geological expertise and experience, the types and severity of water hazards in different regions can be determined. For example, clustered areas near faults with a high probability of hidden dangers may be classified as fault water hazards, with a potentially high degree of hazard. Meanwhile, clustered areas located within aquifers with a relatively low probability of hidden dangers may be classified as general aquifer water hazards, with a relatively low degree of hazard. Accurately determining the types and severity of water hazards in different regions provides a scientific basis for coal mining companies to formulate targeted water hazard prevention and control measures, effectively reducing the risk of water hazard accidents and ensuring safe coal mine production.

[0039] The above-mentioned method for detecting water hazards in coal mine working faces involves constructing an initial geological model containing the three-dimensional distribution of aquifers by acquiring geological structures, lithologic characteristics, and drilling data; identifying the spatial locations of faults and aquifers and obtaining permeability coefficients through geophysical methods; integrating water level dynamic monitoring data to determine the distribution characteristics of old empty water; constructing a prediction model based on the initial model combined with permeability coefficients and old empty water parameters, and outputting a multi-category hidden danger probability distribution map; generating identification results for high-risk areas for water hazards through spatial correlation analysis; and using clustering to classify water hazard types and hazard levels. Its technical effect is to construct a prediction model that can comprehensively reflect the distribution of water hazard hazards for use in the water hazard detection phase of coal mine working face water prevention and control, thereby achieving multi-dimensional and accurate identification of water hazard hazards in coal mine working faces and dynamic risk grading.

[0040] refer to Figure 2 In an optional embodiment, a water hazard distribution prediction model is used to classify and predict water hazard hazards to obtain a multi-category hazard probability distribution map, including the following steps:

[0041] S41: normalizing the fault water permeability coefficient, the aquifer water permeability coefficient, and the old empty water characteristic parameters to obtain normalized parameters; and generating a comprehensive characteristic parameter matrix based on the normalized parameters.

[0042] Specifically, in flood hazard prediction, data such as the permeability coefficient of fault water, the permeability coefficient of aquifer water, and characteristic parameters of old water often have different dimensions and numerical ranges. For example, the permeability coefficient may vary over a large numerical range, while certain characteristic parameters of old water may be relatively small. To eliminate the impact of these differences on subsequent model predictions, the above parameters can be normalized and converted to the same numerical range, usually the interval [0,1]. This method ensures that during the model training and prediction process, the influence of each parameter on the results is balanced, avoiding situations where the model is overly sensitive to certain parameters or ignores other parameters due to differences in numerical ranges.

[0043] The normalized fault water permeability coefficient, aquifer water permeability coefficient, and old water characteristic parameters are arranged in a specific order to form a comprehensive characteristic parameter matrix. Each row of this matrix represents a sample point (such as a grid cell in the working face), and each column represents a characteristic parameter (such as the fault water permeability coefficient, aquifer water permeability coefficient, and the distribution range of old water). The comprehensive characteristic parameter matrix integrates these various parameters, providing a comprehensive data foundation for subsequent model input. This allows the model to simultaneously consider the impact of multiple factors on water hazard risks, improving the accuracy and reliability of predictions.

[0044] S42: Input the comprehensive characteristic parameter matrix into the water hazard distribution prediction model, and output the water hazard type label set Y = {old water, fault water, aquifer water} for each grid cell in the initial geological model; wherein, the probability value P∈[0,1] corresponding to each element of the water hazard type label set Y; the water hazard distribution prediction model is a pre-trained random forest classifier.

[0045] Specifically, the flood hazard distribution prediction model used in this step is a pre-trained random forest classifier. Random forest is an ensemble learning method that constructs multiple decision tree classifiers and votes or averages their prediction results to obtain the final prediction result. Random forest has strong generalization capabilities, can handle high-dimensional data and nonlinear relationships, and has good adaptability to complex flood hazard classification problems. During the model pre-training stage, a large amount of historical data and known flood hazard samples have been used to train the model, allowing the model to learn the intrinsic connections and laws between different feature parameters and flood hazard types.

[0046] The comprehensive feature parameter matrix generated in step S41 is input into the pre-trained random forest classifier. The model classifies and predicts each sample point (grid unit) based on its internal multiple decision trees. Each decision tree will classify the sample point according to its own rules and feature selection, and then vote or weighted average the classification results of all decision trees to obtain the final prediction result. For each grid unit, the model outputs a water hazard type label set Y = {old empty water, fault water, aquifer water}, where the probability value P∈[0,1] corresponding to each element represents the probability that the grid unit belongs to the corresponding water hazard type. For example, if the prediction result of a grid unit is Y = {old empty water: 0.6, fault water: 0.3, aquifer water: 0.1}, it means that the grid unit has a 60% probability of belonging to the old empty water hazard, a 30% probability of belonging to the fault water hazard, and a 10% probability of belonging to the aquifer water hazard.

[0047] These probability values ​​reflect the relative likelihood of each grid cell belonging to a different flood hazard type, providing an important basis for subsequent flood hazard assessment and decision-making. Probability values ​​provide an intuitive understanding of the distribution of flood hazards within the working area and the relative risk levels of different types of flood hazards. For example, when identifying high-risk areas, the probability values ​​can be used to determine high-risk areas. When formulating flood control measures, the probability distribution of different types of flood hazards can be used to tailor appropriate measures, improving prevention effectiveness and resource utilization efficiency.

[0048] S43: Generate a multi-category hidden danger probability distribution map based on the water hazard type label set Y.

[0049] Specifically, according to the water hazard type label set Y and its corresponding probability value of each grid cell obtained in step S42, a multi-category hidden danger probability distribution map is generated using visualization technologies such as geographic information system (GIS). In GIS software, the working face area can be divided into multiple grid cells, and each grid cell corresponds to a location point. According to the water hazard type probability value of each grid cell, the location point is assigned a corresponding color or symbol to indicate the risk level of different water hazard types. For example, red can be used to indicate high-risk areas for old empty water hazards, yellow to indicate high-risk areas for fault water hazards, and green to indicate high-risk areas for aquifer water hazards. The depth of the color or the size of the symbol can indicate the size of the probability value, thereby intuitively displaying the probability distribution of different water hazard hazards in the working face area.

[0050] Multi-category hazard probability distribution maps provide intuitive and visual decision-making support for flood control efforts within coal mine working faces. This map allows managers and technicians to clearly understand the types and risk levels of flood hazards at different locations within the working face, enabling them to develop targeted flood control measures and emergency response plans. For example, drainage and grouting measures can be strengthened in high-risk areas, while resources can be appropriately reduced in low-risk areas, improving the efficiency and effectiveness of prevention efforts. Furthermore, this map can serve as a basis for oversight and inspection by coal mine safety regulators, providing strong technical support for ensuring coal mine production safety.

[0051] In an optional embodiment, S4 further includes the following steps:

[0052] Determine abnormal grid cells from all grid cells using preset rules;

[0053] When the proportion of abnormal grid cells exceeds a preset value, step S1 is re-executed to update the initial geological model, and the comprehensive characteristic parameter matrix is ​​updated based on the updated initial geological model.

[0054] Specifically, the preset rules can be formulated based on a variety of factors, including the geological laws of flood hazard distribution, historical flood hazard data, and the characteristics of the flood hazard probability distribution map. For example, if the flood hazard probability value of a certain grid unit is significantly higher than that of the surrounding grid units, and is inconsistent with the known geological structure, lithological characteristics, etc., or its probability value changes dramatically in a short period of time, then the grid unit may be judged as an abnormal grid unit. In addition, corresponding rules can be set based on the degree of difference in the probability values ​​of each element in the flood hazard type label set, the change trend, etc., to identify grid units that may have abnormalities.

[0055] By analyzing the probability distribution of water hazards across all grid cells, each grid cell is checked individually according to pre-set rules to see if it meets the abnormality criteria. This step can involve statistical analysis of the probability values ​​of the grid cells, such as calculating statistics such as the mean and standard deviation, and then setting an abnormality threshold based on these statistics; or using anomaly detection algorithms from data mining and machine learning, such as isolation forests and local anomaly factors, to automatically identify abnormal grid cells. If a grid cell is found that meets the abnormality criteria, it is identified as an abnormal grid cell, and its location, probability value, and other information are recorded.

[0056] The preset value can be set based on factors such as the geological conditions of the coal mine working face, the distribution characteristics of water hazard hazards, and the requirements for water hazard prediction accuracy. It represents a threshold value for the proportion of abnormal grid cells in all grid cells. When the proportion of abnormal grid cells exceeds the preset value, it indicates that the current initial geological model may have large deviations or inaccuracies and needs to be updated and corrected. For example, if the preset value is set to 10%, when it is found that the proportion of abnormal grid cells in all grid cells exceeds 10%, the mechanism of re-executing step S1 is triggered.

[0057] The purpose of re-executing step S1 is to update the initial geological model based on the latest geological data and flood hazard information. During the coal mining process, geological conditions may change, such as the discovery of new geological structures and changes in existing lithologic characteristics. These changes may affect the distribution of flood hazards. By re-collecting and organizing working face structural data, lithologic characteristics data, and drilling data, and constructing a new initial geological model, we can more accurately reflect the current geological conditions and provide a more reliable basis for subsequent flood hazard predictions. At the same time, updating the initial geological model also helps improve the accuracy and reliability of flood hazard predictions and reduce the risk of flood accidents.

[0058] Updating the initial geological model will cause changes in related geological parameters, such as fault location, aquifer distribution, and lithologic characteristics. The comprehensive characteristic parameter matrix is ​​generated based on these parameters in the initial geological model, so it can be updated accordingly based on the updated initial geological model. This ensures that the data in the comprehensive characteristic parameter matrix accurately reflects the current geological conditions and water hazard characteristics, providing accurate input data for subsequent water hazard classification and prediction.

[0059] According to the updated initial geological model, the characteristic parameters related to water hazards are re-extracted, such as the fault water permeability coefficient, the aquifer water permeability coefficient, the old empty water characteristic parameters, etc. Then, according to the normalization processing method in step S41, the updated characteristic parameters are normalized to generate a new comprehensive characteristic parameter matrix. During the updating process, it is important to keep the format and structure of the comprehensive characteristic parameter matrix consistent with the previous one, so that it can be smoothly input into the water hazard distribution prediction model for subsequent prediction analysis. By continuously updating the comprehensive characteristic parameter matrix, the water hazard prediction model can adapt to changes in geological conditions in a timely manner, thereby improving the accuracy and timeliness of the prediction results.

[0060] In an optional embodiment, S2 includes the following steps:

[0061] S21: Based on the fault strike information in the working face structural data, a combined detection technology of resistivity method and electromagnetic wave reflection method is used to obtain fault strike dip data and aquifer top and bottom elevation data.

[0062] Specifically, the resistivity method is a geophysical prospecting method based on the fact that different rocks and geological bodies have different resistivity characteristics. By arranging electrodes on the ground or in boreholes, supplying power to the underground, and measuring the underground resistivity distribution, the method is used. In the detection of water prevention and control at coal mine working faces, the electrode array is reasonably arranged according to the fault strike information in the working face structural data, and measurements are performed along the fault strike and dip direction. The resistivity data obtained by measurement can reflect the resistivity differences between different underground rock layers and geological bodies, and then infer the location, strike and dip of the fault, and the distribution of the aquifer. During the implementation process, appropriate parameters such as electrode spacing and power supply current are selected to ensure the accuracy and reliability of the measurement data.

[0063] The electromagnetic wave reflection method exploits the property of artificially excited electromagnetic waves, which reflect when propagating through underground media and encounter different electrical interfaces. This method detects underground geological structures by receiving and analyzing the waveform, amplitude, phase, and other information of the reflected waves. In coal mine working face detection, electromagnetic wave reflection detection is also performed within the working face area, using appropriate electromagnetic wave frequencies and transmission powers based on fault strike information. When electromagnetic waves encounter electrical interfaces such as fault interfaces or aquifers, they generate reflected waves. Analysis of these reflected waves can determine the spatial location of the fault, its strike and dip, and the top and bottom elevations of the aquifer. In actual operation, appropriate electromagnetic wave frequencies, transmission powers, and reception parameters are selected based on the geological conditions and the purpose of the detection to improve detection resolution and accuracy.

[0064] Combining resistivity and electromagnetic wave reflection methods can leverage the strengths of both methods and overcome the shortcomings of each alone. The resistivity method is effective for detecting aquifers, visually reflecting their resistivity characteristics, but its ability to detect fine fault structures is relatively limited. Electromagnetic wave reflection methods, on the other hand, offer high resolution for fault detection, accurately determining their spatial location and strike and dip angles. However, their effectiveness in detecting aquifers may be affected by electromagnetic wave attenuation within the aquifer. By combining these detection techniques, detailed information on both faults and aquifers can be obtained simultaneously, providing a more accurate data foundation for subsequent water hazard prediction.

[0065] S22: Generate fault spatial position data based on the fault strike and dip data; the fault spatial position data includes the coordinates of the fault center point (x f ,y f ,z f ) and extension length L f ; where x f and y f is the horizontal coordinate, z f is the vertical depth coordinate.

[0066] Specifically, by processing and analyzing the large amount of raw data obtained from the resistivity and electromagnetic wave reflection methods, characteristic information related to faults and aquifers is extracted. For resistivity data, data processing methods such as inversion algorithms are used to convert the measured resistivity data into underground resistivity distribution images. Based on the abnormal resistivity changes in the images, the location of the aquifer and the top and bottom plate elevations are determined. For electromagnetic wave reflection data, seismic data processing techniques such as filtering and offset imaging are used to process the reflected wave data to obtain a reflected wave profile of the underground geological structure. Based on the characteristic changes in the reflected wave phase axis, the spatial position and strike angle of the fault are determined.

[0067] Based on the extracted fault strike and dip data, combined with the working surface coordinate system, the fault spatial position data is generated. The fault spatial position data includes the coordinates of the fault center point (x f ,y f ,z f ) and extension length L f The coordinates of the center point of the fault can be calculated by fitting the position of the fault in three-dimensional space, and the extension length L f The data is determined based on the actual extension of the fault within the working area. The above data can accurately describe the position and morphological characteristics of the fault in the underground space, providing key geometric parameters for subsequent water hazard prediction and risk assessment.

[0068] S23: Integrate the aquifer top and bottom elevation data to obtain the aquifer spatial position data, which includes the aquifer top elevation Htop and aquifer bottom elevation H bot .

[0069] Specifically, the aquifer top and bottom elevation data obtained by the resistivity method and the electromagnetic wave reflection method are integrated to obtain the aquifer spatial position data. The aquifer spatial position data includes the aquifer top elevation H top and aquifer bottom elevation H bot By interpolating and fitting the aquifer top and bottom elevation data at different locations, a three-dimensional spatial distribution model of the aquifer within the working surface area is generated, which can intuitively display the top and bottom morphology and spatial position relationship of the aquifer. During the data integration process, the accuracy and reliability of the data are taken into consideration, and data with potential errors are corrected and processed to ensure the accuracy and completeness of the aquifer spatial location data.

[0070] Aquifer spatial location data is crucial for coal mine working face water prevention and control. Accurately understanding the top and bottom elevations and spatial distribution of aquifers allows for the determination of the spatial relationship between the aquifer and the working face, assessing the aquifer's impact on working face mining. Furthermore, aquifer spatial location data is crucial for developing models predicting the distribution of water hazards. It provides accurate aquifer boundary conditions and hydrogeological parameters, improving the accuracy and reliability of these models.

[0071] S24: Obtain the fault water inflow Q through pumping test f and fault hydraulic conductivity T f Calculate the fault water permeability coefficient K according to the following Theis formula f :

[0072]

[0073] Where W(u) is the well function, and the value is selected from the standard well function table according to the lithologic type of the fault zone; u = r 2 S / (4T f t); r is the horizontal distance between the observation hole and the pumping hole in the coal mine working face; S is the water storage coefficient of the fault zone, which is calculated by the water release corresponding to the unit water level drop in the pumping test; t is the duration of the pumping test.

[0074] Specifically, pumping holes are arranged near the coal mine working face or near the fault, and a pumping test is conducted. The design of the pumping test can take into account factors such as the location, depth, and diameter of the pumping hole, as well as the selection of pumping equipment, to ensure that the test can accurately obtain the hydrogeological parameters of the fault water. During the pumping test, water level gauges can be installed in the pumping hole and surrounding observation holes to monitor water level changes in real time and record the amount of water pumped during the pumping process. Through long-term pumping tests, water level drop and pumping volume data can be obtained at different time stages.

[0075] According to the water level drop and pumping volume data recorded in the pumping test, the relevant formulas and methods in hydrogeology are used to calculate the inflow volume Q of the fault water. f and hydraulic conductivity T f Fault water inflow Q f It can be obtained by directly measuring the amount of water pumped during the pumping process, and the hydraulic conductivity T f The calculation can be done based on the relationship between water level drop and pumping volume. f It is an important parameter reflecting the water conductivity of the fault. It is related to factors such as the lithology, structure, and degree of fracture development of the fault. The value of this parameter can be accurately obtained through pumping tests.

[0076] Calculate the fault water permeability coefficient K using the Theis formula f , the formula is Where W(u) is the well function, and the value can be selected from the standard well function table according to the lithologic type of the fault zone; u = r 2 S / (4T f t); r is the horizontal distance between the observation hole and the pumping hole in the coal mine working face; S is the fault zone water storage coefficient, which can be calculated by the water release corresponding to the unit water level drop in the pumping test; t is the duration of the pumping test. By accurately calculating the above parameters and substituting them into the formula, the fault water permeability coefficient K can be obtained f The numerical value of .

[0077] S25: Obtain the water yield Q of the aquifer through water injection test a and aquifer conductivity T a Calculate the water permeability coefficient K of the aquifer according to the following formula a :

[0078]

[0079] Among them, H a =H top -H bot ;L a is the length of the injection section in the water injection test, determined according to the elevation of the aquifer top; r ais the radius of the water injection hole in the water injection test.

[0080] Specifically, injection holes are placed in the aquifer and a water injection test is conducted. The design of the water injection test can take into account factors such as the location, depth, and diameter of the injection holes, as well as the selection of injection equipment, to ensure that the test accurately captures the hydrogeological parameters of the aquifer. During the water injection test, pressure sensors and flow meters can be installed in the injection holes to monitor changes in injection pressure and injection volume in real time and record relevant data during the injection process. Through long-term water injection tests, injection pressure and injection volume data can be obtained at different time stages.

[0081] According to the injection pressure and injection volume data recorded in the water injection test, the water inflow volume Q of the aquifer water is calculated using the relevant formulas and methods in hydrogeology. a and hydraulic conductivity T a Aquifer water inflow Q a It can be obtained by directly measuring the water injection volume during the water injection process, and the hydraulic conductivity T a The calculation can be done based on the relationship between injection pressure and injection volume. a It is an important parameter reflecting the water conductivity of the aquifer. It is related to factors such as the lithology, pore structure, and permeability of the aquifer. The value of this parameter can be accurately obtained through water injection tests.

[0082] According to the formula Calculate the water permeability coefficient K of the aquifer a Among them, H a =H top -H bot ;L a The length of the injection section in the water injection test can be determined according to the elevation of the aquifer top; r a is the radius of the injection hole in the water injection test. By accurately calculating the above parameters and substituting them into the formula, the water permeability coefficient K of the aquifer can be obtained. a The numerical value of .

[0083] In an optional embodiment, S3 includes the following steps:

[0084] S31: Obtain a monitoring data set including a water level change rate and a daily water inflow fluctuation value from the water level dynamic monitoring data.

[0085] Specifically, the dynamic monitoring data of water levels can come from water level monitoring points set up around the coal mine working face and the old empty area. The monitoring point is equipped with a high-precision water level sensor that can collect information on changes in groundwater levels in real time or regularly. The monitoring data content includes but is not limited to the real-time height of the water level, the time series of water level changes, etc. By analyzing the above data, the water level change rate, that is, the speed at which the water level rises or falls per unit time, and the daily water inflow fluctuation value, that is, the maximum change in water inflow within a day, can be obtained. The above data provide important basic information for the subsequent old empty water recharge path simulation, and can reflect the dynamic change characteristics of old empty water in the time dimension.

[0086] The collected dynamic water level monitoring data is collated and filtered to construct a monitoring dataset that includes water level change rates and daily water inflow fluctuations. During this construction process, the data can be preprocessed, such as removing outliers and filling missing values, to ensure data accuracy and completeness. Furthermore, the data can be appropriately formatted and normalized based on actual needs and the requirements of the simulation software, making it recognizable and usable by hydrogeological numerical simulation software. The construction of this monitoring dataset provides direct data support for subsequent simulation analysis, enabling simulation results to more accurately reflect the actual situation of the old empty water.

[0087] S32: Input the monitoring data set and regional hydrogeological condition data into the hydrogeological numerical simulation software to generate the simulation results of the old empty water recharge path.

[0088] Specifically, regional hydrogeological data includes information on the geological structure, lithologic characteristics, groundwater recharge sources, groundwater runoff conditions, and groundwater discharge conditions of the old, empty areas. This data can come from geological exploration reports, hydrogeological survey data, and relevant Geographic Information System (GIS) data. Before simulation, this data can be integrated and processed to have the same spatial reference and data format as the monitoring dataset, making it usable by hydrogeological numerical simulation software.

[0089] Selecting appropriate hydrogeological numerical simulation software is the key to simulating the recharge path of old empty water. The hydrogeological numerical simulation software that can be used include MODFLOW, FEFLOW, etc. The above software has powerful groundwater simulation functions and can simulate complex processes such as groundwater flow, solute migration, and multiphase flow. In this step, the monitoring data set and regional hydrogeological condition data are input into the selected hydrogeological numerical simulation software, the corresponding simulation parameters and boundary conditions are set, the simulation program is run, and the simulation results of the recharge path of old empty water are generated. The simulation results can be presented in the form of a three-dimensional visual model or a two-dimensional profile diagram, which can intuitively show the recharge source, recharge path and distribution of old empty water in the underground space.

[0090] S33: Extract features from the simulation results of the old empty water supply path to obtain characteristic parameters of the old empty water.

[0091] Specifically, feature extraction is performed on the simulation results of the old empty water recharge path, mainly to extract parameters related to the old empty water characteristics from the simulation results. Such parameters include but are not limited to the distribution range of the old empty water, the source of recharge, and the water quality characteristics. The distribution range can be determined by the groundwater isowater level lines or isoconcentration lines in the simulation results, the source of recharge can be identified by analyzing the groundwater flow and velocity vector field in the simulation results, and the water quality characteristics can be evaluated by the solute concentration distribution in the simulation results. The feature extraction method can adopt image processing technology, numerical analysis methods, and professional hydrogeological analysis software, etc., and the required characteristic parameters can be extracted by processing and analyzing the simulation results.

[0092] The obtained characteristic parameters of old goaf water provide an important basis for water prevention and control efforts in coal mine working faces. The distribution range parameters can help determine the working face areas potentially affected by old goaf water, providing a spatial scope for identifying and warning of water hazards. The recharge source parameters reveal the recharge mechanism of old goaf water, providing guidance for prevention and control measures such as cutting off recharge pathways and lowering old goaf water levels. The water quality characteristic parameters reflect the water quality of old goaf water, providing a reference for developing appropriate water quality treatment plans and environmental protection measures. Through the comprehensive analysis and application of the characteristic parameters of old goaf water, the scientific nature and effectiveness of water prevention and control efforts in coal mine working faces can be effectively improved, reducing the risk of water hazards.

[0093] In an optional embodiment, S3 further includes the following steps:

[0094] When the water level change rate exceeds the first threshold or the daily water inflow fluctuation value exceeds the second threshold, step S2 is re-executed to update the fault water permeability coefficient and the aquifer water permeability coefficient.

[0095] Specifically, the first and second thresholds can be set based on a review of experience gained in coal mine working face flood control, as well as statistical analysis of dynamic water level monitoring data and daily water inflow fluctuations. By analyzing extensive historical data, the appropriate ranges for the rate of water level change and daily water inflow fluctuations under normal circumstances are determined. Then, appropriate thresholds are set based on a safety factor and early warning mechanism. These thresholds are designed to promptly detect abnormal changes in water levels and water inflow, allowing appropriate measures to be taken to prevent flooding accidents.

[0096] When analyzing dynamic water level monitoring data, the water level change rate and daily water inflow fluctuation are calculated in real time or periodically and compared with predefined first and second thresholds. If the water level change rate exceeds the first threshold, or the daily water inflow fluctuation exceeds the second threshold, the monitoring data is considered abnormal. This may indicate a change in the recharge conditions of the depleted water supply or abnormal groundwater flow caused by geological factors such as tectonic structures. Further updates and assessments of relevant hydrogeological parameters are necessary.

[0097] When abnormalities occur in the monitoring data, the original fault water permeability coefficient and aquifer water permeability coefficient may no longer accurately reflect the current hydrogeological conditions. Because the above parameters are obtained through methods such as pumping tests or water injection tests under certain geological and hydrological conditions, when the conditions change, the parameter values ​​will also change accordingly. If the above parameters are not updated in a timely manner, the accuracy of subsequent water hazard predictions will be affected, which may lead to misjudgment or omission of water hazard hazards, thereby posing a safety hazard to coal mine production. Therefore, step S2 can be re-executed to obtain the latest fault water permeability coefficient and aquifer water permeability coefficient to ensure the reliability and accuracy of water hazard predictions.

[0098] When re-executing step S2, the combined detection technology of resistivity method and electromagnetic wave reflection method is used again to obtain the fault strike and dip data and the aquifer top and bottom plate elevation data, and the fault water inflow and water conductivity, aquifer water inflow and water conductivity are obtained respectively through pumping test and water injection test. Then, according to the Theis formula and related calculation formulas, the fault water permeability coefficient and the aquifer water permeability coefficient are recalculated. During the calculation process, attention should be paid to the standardization of the test method and the accuracy of the data to ensure that the updated parameters can truly reflect the current hydrogeological conditions. At the same time, the updated parameters should be reasonably analyzed and interpreted, and combined with factors such as geological structure and lithological characteristics, the degree of their impact on water hazard risks should be evaluated to provide a scientific basis for subsequent water hazard prevention and control work.

[0099] In an optional embodiment, S5 includes the following steps:

[0100] S51: According to the extension length L in the slice space position data f and vertical depth coordinate z f , using the formula Calculate the fault impact weight value W f .

[0101] Specifically, L f Indicates the extension length of the fault, reflecting the spatial distribution range of the fault; z f is the vertical depth coordinate of the fault, indicating the position of the fault in the vertical direction; H topis the elevation of the aquifer top. This formula calculates the fault's impact weight on water hazard risk by comparing its extension length to its vertical distance above the aquifer. The longer the extension length, the larger the potential impact range of the fault; and the smaller the difference between the vertical depth coordinate and the aquifer top elevation, the closer the fault is to the aquifer and the more direct its impact on the aquifer. Therefore, the fault impact weight W is f The bigger it is.

[0102] Calculate the fault impact weight value W f The purpose of this is to quantify the contribution of faults to water hazard risk. In coal mine working faces, the presence of faults may cause groundwater to infiltrate along the fault fissures, increasing the possibility of water hazard. By calculating W f , we can clarify the impact of different faults on water hazard risk, and provide an important basis for subsequent water hazard risk assessment and regional division. For example, in the case of multiple faults, we can f Sort the faults by size, give priority to those with greater impact on water disaster risks, and take corresponding prevention and control measures.

[0103] S52: According to the aquifer top elevation H in the aquifer spatial position data top and aquifer bottom elevation H bot , using the formula W a =(H top -H bot )×K a Calculate the aquifer water richness index W a .

[0104] Specifically, according to the aquifer top elevation H in the aquifer spatial position data top and aquifer bottom elevation H bot , using the formula W a =(H top -H bot )×K a Calculate the aquifer water richness index W a Among them, (H top -H bot ) represents the thickness of the aquifer, reflecting the size of the water storage space of the aquifer; K a The coefficient of water permeability of the aquifer indicates its ability to penetrate groundwater. This formula multiplies the aquifer thickness by the permeability coefficient to produce the aquifer's water-richness index, which comprehensively reflects the aquifer's water storage capacity and permeability.

[0105] Aquifer water richness index W aIt is an important indicator for assessing the contribution of aquifers to water hazard risks. The thicker the aquifer, the larger the water storage space and the more water it can store; the larger the permeability coefficient, the stronger the aquifer's ability to penetrate groundwater, making it easier for groundwater to penetrate through the aquifer into the coal mine working face. By calculating W a , can quantify the water richness of aquifers and provide key parameters for water hazard risk assessment. For example, in the case of coexistence of multiple aquifers, the water richness of aquifers can be quantified based on W a The size of the aquifer with stronger water-yield is used to determine the aquifer with stronger water-yield, and attention is paid to the impact of the aquifer with stronger water-yield on the risk of water disasters, and corresponding prevention and control measures are taken.

[0106] S53: Combine the multi-category hidden danger probability distribution map with the fault impact weight value W f , aquifer water richness index W a Perform spatial weighted fusion to generate a water hazard risk index distribution map.

[0107] Specifically, the probability distribution map of multi-category hidden dangers and the fault impact weight value W f , aquifer water richness index W a Specifically, the hidden danger probability value of each grid unit can be combined with the corresponding W f and W a Multiplying the values ​​by 0 and 1 yields a weighted hazard probability value. This weighted hazard probability value is then normalized to lie within the [0, 1] interval, generating a flood risk index distribution map. The principle of spatial weighted fusion is to adjust the hazard probability distribution map based on the impact of faults and aquifers on flood risk, ensuring that the flood risk index distribution map more accurately reflects the actual flood risk situation.

[0108] By generating a water hazard risk index distribution map through spatial weighted fusion, we can comprehensively consider the impact of multiple factors such as faults, aquifers, and hidden danger probability on water hazard risk. The multi-category hidden danger probability distribution map reflects the spatial probability distribution of different water hazard hidden danger types, but does not fully consider the impact of the characteristics of faults and aquifers on water hazard risk. The fault impact weight value W f and aquifer water richness index W a The contribution of faults and aquifers to water hazard risk is quantified separately. By integrating these factors through spatial weighted fusion, the resulting water hazard risk index distribution map can more comprehensively and accurately reflect the water hazard risk status of coal mine working faces, providing a more reliable basis for subsequent regional division and the formulation of prevention and control measures.

[0109] S54: Based on the preset risk threshold, the water hazard risk index distribution map is divided into regions to obtain high-risk areas, medium-risk areas and low-risk areas, and the high-risk area is used as the identification result.

[0110] Specifically, preset risk thresholds are set based on the actual conditions of the coal mine working face and flood control requirements. These thresholds can be categorized as high, medium, and low, and are used to divide the flood risk index distribution map into high-risk, medium-risk, and low-risk areas. The risk thresholds can be set by comprehensively considering factors such as the coal mine's geological conditions, mining processes, and historical flood records, ensuring that the resulting classifications accurately reflect the actual flood risk situation.

[0111] The flood risk index distribution map is divided into regions based on preset risk thresholds. Specifically, areas with a flood risk index greater than the high-risk threshold are classified as high-risk zones, areas with a flood risk index between the medium-risk and high-risk thresholds are classified as medium-risk zones, and areas with a flood risk index less than the medium-risk threshold are classified as low-risk zones. This regional division can clarify the flood risk level of different areas and provide a basis for taking targeted prevention and control measures.

[0112] High-risk areas are identified as key areas for flood control efforts. More stringent prevention and control measures can be implemented in these areas, such as enhanced drainage, grouting to block water, and preemptive water exploration and release, to reduce the risk of flooding. For medium- and low-risk areas, appropriate prevention and control measures can be implemented based on actual conditions, rationally allocating resources and improving the efficiency and effectiveness of flood control. By developing prevention and control measures based on regional zoning, the risk of flooding at coal mine working faces can be effectively reduced, ensuring safe production in coal mines.

[0113] In an optional embodiment, S1 includes the following steps:

[0114] S11: Based on the aquifer depth and thickness in the drilling data, the Kriging interpolation method is used to generate the three-dimensional aquifer top and bottom surface.

[0115] Specifically, Kriging interpolation is a geostatistical interpolation method that fully considers the spatial correlation and variability of geological variables. It describes the spatial correlation between geological variables by constructing a variogram model, thereby achieving optimal unbiased estimation of geological variables at unknown points. Compared with traditional interpolation methods, Kriging interpolation can better reflect the spatial distribution of geological variables, and the generated interpolation results are more accurate and reliable.

[0116] Based on the aquifer depth and thickness information from the drilling data, the elevation data for the aquifer roof and floor at each borehole location are determined. Kriging interpolation is then used to spatially interpolate the elevation data to generate three-dimensional aquifer roof and floor surfaces. During the interpolation process, a variogram analysis is first performed on the data to determine the parameters of the variogram model, such as the sill, nugget value, and range. Interpolation calculations are then performed based on the variogram model. The resulting three-dimensional aquifer roof and floor surfaces intuitively demonstrate the spatial morphology and distribution characteristics of the aquifer.

[0117] S12: Encode the lithologic characteristic data to obtain rock permeability classification parameters.

[0118] Specifically, lithologic characteristic data includes information such as rock type, structure, and composition, which significantly influences rock permeability. To convert lithologic characteristic data into parameters useful for quantitative analysis, they are encoded. Encoding methods can employ numerical or symbolic coding, assigning different numerical values ​​or symbols to different lithologic characteristics to indicate their impact on rock permeability. For example, lithologic characteristic data can be categorized by rock type, such as sandstone, mudstone, and limestone, with each category assigned a corresponding numerical code. Similar encoding methods can also be used for rock characteristics such as structure and composition.

[0119] The resulting rock permeability classification parameters reflect the differences in permeability of rocks with different lithologic characteristics, providing an important basis for assessing the permeability of groundwater within rock formations and the risk of water hazards. By combining these rock permeability classification parameters with spatial location data within the initial geological model, the permeability distribution of different rock formations can be visually displayed, providing a scientific basis for subsequent water hazard prediction and prevention measures. For example, in areas with high permeability, groundwater flow may be more active, the risk of water hazards is relatively high, and more stringent prevention and control measures can be implemented.

[0120] S13: Spatially align the three-dimensional aquifer top and bottom surface, rock permeability classification parameters, and working face structural data with the fault spatial position data and the aquifer spatial position data to generate an initial geological model.

[0121] Specifically, spatial registration is the process of aligning and integrating data from different sources within a unified spatial coordinate system. In this step, the three-dimensional aquifer top and bottom surface, rock permeability classification parameters, and working face structural data are spatially registered with the fault spatial position data and aquifer spatial position data. First, a unified spatial coordinate system is determined, such as a national coordinate system or a mining area's own coordinate system. Then, based on the data's geographic location information, the various data sets are aligned and integrated. This step may involve operations such as coordinate conversion and projection transformation to ensure accurate spatial correspondence between all data.

[0122] Through spatial registration, all relevant data are integrated into a unified spatial framework to generate an initial geological model. This model contains important information such as aquifer distribution, rock permeability, and geological structure within the working face area, providing a comprehensive and accurate geological foundation for subsequent water hazard prediction and prevention efforts. Analysis and research of the initial geological model provides a deep understanding of the geological conditions and hydrogeological characteristics of the working face area, providing strong support for the development of scientific and rational water hazard prevention and control measures. Furthermore, the initial geological model serves as the foundational data for subsequent numerical simulations and predictive analysis, ensuring the accuracy and reliability of water hazard predictions.

[0123] The above-mentioned water hazard detection method for coal mine working face water prevention and control integrates multi-source information such as geological structure, lithological characteristics, drilling data and hydrogeological conditions, and uses Kriging interpolation method, random forest classifier, hydrogeological numerical simulation and spatial weighted fusion and other technical means to construct and update the initial geological model, generate multi-category hidden danger probability distribution map and water hazard risk index distribution map, and ultimately achieve high-precision prediction of water hazard hazards in coal mine working faces, risk level classification and accurate identification of high-risk areas, effectively improve the scientificity and accuracy of coal mine water hazard prevention and control, and provide strong technical guarantee for coal mine safety production.

[0124] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0125] An embodiment of the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0126] An embodiment of the present application further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0127] The above-described embodiments merely represent several implementation methods of the embodiments of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that a person skilled in the art may make various modifications and improvements without departing from the concept of the embodiments of the present application, and these modifications and improvements fall within the scope of protection of the embodiments of the present application.

Claims

1. A method for detecting water hazards in coal mine working faces, characterized in that: The method comprises: S1: constructing an initial geological model based on working face structural data, lithologic characteristic data, and drilling data, wherein the initial geological model includes three-dimensional spatial distribution information of the aquifer; S2: Based on the working face structural data, a geophysical method is used to obtain fault spatial position data and aquifer spatial position data, and to obtain the fault water permeability coefficient and the aquifer water permeability coefficient; S3: Obtaining characteristic parameters of old empty water based on water level dynamic monitoring data and regional hydrogeological conditions data; the characteristic parameters of old empty water include distribution range, supply source and water quality characteristics; S4: Based on the initial geological model, in combination with the fault spatial position data, the aquifer spatial position data, the fault water permeability coefficient, the aquifer water permeability coefficient, and the old empty water characteristic parameters, a water hazard distribution prediction model is constructed; and the water hazard distribution prediction model is used to classify and predict water hazard hazards to obtain a multi-category hazard probability distribution map; S5: performing spatial correlation analysis on the multi-category hidden danger probability distribution map, the fault spatial location data, and the aquifer spatial location data to generate an identification result of a high-risk flood disaster area; S6: Based on the identification results, cluster the high-risk areas for flood disasters to obtain clustering results; and based on the clustering results, determine the types and levels of flood disasters in different areas.

2. The method according to claim 1, characterized in that The water hazard distribution prediction model is used to classify and predict water hazard hazards to obtain a multi-category hazard probability distribution map, including: Normalizing the fault water permeability coefficient, the aquifer water permeability coefficient, and the old empty water characteristic parameters to obtain normalized parameters; generating a comprehensive characteristic parameter matrix based on the normalized parameters; Inputting the comprehensive characteristic parameter matrix into the water hazard distribution prediction model, outputting a water hazard type label set Y = {old water, fault water, aquifer water} for each grid cell in the initial geological model; wherein the probability value P∈[0,1] corresponding to each element of the water hazard type label set Y; the water hazard distribution prediction model is a pre-trained random forest classifier; The multi-category hidden danger probability distribution map is generated according to the water hazard type label set Y.

3. The method according to claim 2, characterized in that Said S4 further comprises: Determining abnormal grid cells from all the grid cells according to preset rules; When the proportion of the abnormal grid cells exceeds a preset value, step S1 is re-executed to update the initial geological model, and the comprehensive characteristic parameter matrix is ​​updated based on the updated initial geological model.

4. The method according to claim 1, wherein The S2 includes: Based on the fault strike information in the working face structural data, a combined detection technology of resistivity method and electromagnetic wave reflection method is used to obtain fault strike dip data and aquifer top and bottom elevation data; The fault spatial position data is generated according to the fault strike and dip data; the fault spatial position data includes the coordinates of the fault center point (x f ,y f ,z f ) and extension length L f ; where x f and y f is the horizontal coordinate, z f is the vertical depth coordinate; Integrate the aquifer top and bottom elevation data to obtain the aquifer spatial position data, which includes the aquifer top elevation H top and aquifer bottom elevation H bot ; Obtaining the fault water inflow Q through pumping test f and fault hydraulic conductivity T f Calculate the fault water permeability coefficient K according to the following Theis formula: f : Where W(u) is the well function, and the value is selected from the standard well function table according to the lithologic type of the fault zone; u = r 2 S / (4T f t); r is the horizontal distance between the observation hole and the pumping hole in the coal mine working face; S is the fault zone water storage coefficient, which is calculated by the water release corresponding to the unit water level drop in the pumping test; t is the duration of the pumping test; Obtain the water yield Q of the aquifer through water injection test a and aquifer conductivity T a Calculate the water permeability coefficient K of the aquifer according to the following formula: a : Among them, H a =H top -H bot ;L a is the length of the injection section in the water injection test, which is determined according to the elevation of the aquifer top; r a is the radius of the water injection hole in the water injection test.

5. The method according to claim 1, wherein The S3 includes: Acquire a monitoring data set including a water level change rate and a daily water inflow fluctuation value from the water level dynamic monitoring data; Inputting the monitoring data set and the regional hydrogeological condition data into hydrogeological numerical simulation software to generate simulation results of the old empty water recharge path; Feature extraction is performed on the simulation results of the old empty water replenishment path to obtain the old empty water characteristic parameters.

6. The method according to claim 5, characterized in that Said S3 further comprises: When the water level change rate exceeds a first threshold or the daily water inflow fluctuation value exceeds a second threshold, step S2 is re-executed to update the fault water permeability coefficient and the aquifer water permeability coefficient.

7. The method according to claim 4, characterized in that The S5 includes: According to the extension length L in the tomographic space position data f and the vertical depth coordinate z f , use the following formula to calculate the fault influence weight value W f : According to the aquifer top elevation H in the aquifer spatial position data top and the aquifer bottom elevation H bot , use the following formula to calculate the aquifer water richness index W a : W a =(H top -H bot )×K a ; The multi-category hidden danger probability distribution map and the fault impact weight value W f , the water-richness index W of the aquifer a Perform spatial weighted fusion to generate a water hazard risk index distribution map; The water hazard risk index distribution map is divided into regions based on a preset risk threshold to obtain high-risk areas, medium-risk areas and low-risk areas, and the high-risk areas are used as the identification results.

8. The method according to any one of claims 1 to 7, characterized in that Said S1 comprises: According to the aquifer depth and aquifer thickness in the drilling data, a three-dimensional aquifer top and bottom surface is generated by using a Kriging interpolation method; Encoding the lithologic characteristic data to obtain rock permeability classification parameters; The three-dimensional aquifer top and bottom surface curved surfaces, the rock permeability classification parameters and the working face structural data are spatially aligned with the fault spatial position data and the aquifer spatial position data to generate the initial geological model.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images