Coal mine water disaster water inrush early warning grading method

Abstract

The invention provides a coal mine water disaster water inrush early warning grading method, and relates to the technical field of coal mine safety engineering and intelligent monitoring, and the method comprises the steps: collecting the multi-source basic data of a coal mine area in real time; constructing a first-level index-second-level index layered influence factor set based on a water source condition, a water guide channel and a dynamic factor of coal mine water disaster water inrush, and introducing an index dynamic correlation coefficient; defining an early warning grade-symbol-judgment standard-emergency response four-dimensional judgment set of the coal mine water disaster; respectively calculating the membership degree for the measured value of each secondary index in the multi-source basic data, and constructing a fuzzy relation matrix; dynamically calculating the real-time weight of each level of index according to the mining process and the index dynamic correlation coefficient; and based on the real-time weight and the fuzzy relation matrix, calculating and obtaining a corresponding early warning grade and an emergency response in the four-dimensional evaluation set. According to the method, full-period water inrush risk grading early warning of bottom plate limestone water, top plate sandstone water, goaf water and composite water disasters under complex geological conditions can be achieved.

Description

Technical Field

[0001] This invention relates to the field of coal mine safety engineering and intelligent monitoring technology, and in particular to a method for early warning and classification of water inrush in coal mines. Background Technology

[0002] Coal mine water hazards are one of the core bottlenecks restricting the safe mining of coal resources. According to incomplete statistics, in the past five years, 80% of major coal mine water hazard accidents in my country were related to "unclear water source identification, inaccurate channel detection, and delayed dynamic response." Coal mine water hazards are a major hidden danger to coal mine safety production. At best, they lead to the shutdown of the working face and damage to equipment; at worst, they cause mine flooding, casualties, and irreparable losses. Existing early warning technologies have the following key defects: (1) Insufficient fusion of multi-source data: The isolated use of sensor monitoring data such as water level and pressure without integrating key information such as geological data and dynamic detection leads to "disconnect between static geology and dynamic monitoring"; (2) Lack of timeliness of weights: The evaluation uses fixed weights and does not take into account the changes in the contribution of "water source-channel-dynamic factors" during the mining process (such as the channel weight should be increased when approaching the fault), resulting in poor early warning adaptability; (3) Weak emergency response coordination: The early warning results only output the risk level and do not associate it with specific prevention and control engineering parameters (such as drainage capacity matching value), and the on-site disposal lacks timely and accurate guidance; (4) The fuzzification method is singular: a uniform trapezoidal membership function is used for all indicators, and the fuzzification model is not optimized for "strong nonlinear indicators", resulting in low hierarchical resolution.

[0003] Therefore, there is an urgent need to provide a coal mine water inrush early warning and classification method to solve the problems of "insufficient data fusion, static weights, and weak emergency response" in existing methods, and to achieve accurate identification, dynamic classification, and efficient response to water inrush risks. Summary of the Invention

[0004] To address the aforementioned issues, this invention provides a coal mine water inrush early warning and classification method. Based on the entire chain of "multi-source data fusion → dynamic weight allocation → intelligent fuzzification → emergency linkage," it integrates AHP (Advanced Fuzzy Analytical Processing) and Internet of Things (IoT) technologies to construct an integrated early warning method of "accurate identification, scientific classification, and efficient handling," filling the technological gap in refined early warning of water inrush under complex conditions.

[0005] To achieve the above objectives, the present invention provides a method for early warning and classification of water inrush in coal mines, comprising: Real-time acquisition of multi-source basic data in coal mining areas, including static geological data, dynamic detection data, dynamic geophysical data, and hydrochemical data; Based on the water source conditions, water diversion channels, and dynamic factors of coal mine water inrush, a hierarchical set of influencing factors consisting of primary and secondary indicators is constructed. The definitions, quantification methods, and data sources of each indicator are clarified, and dynamic correlation coefficients of the indicators are introduced. Define a four-dimensional evaluation set for early warning levels, symbols, evaluation criteria, and emergency response to coal mine water hazards; The membership degree is calculated for the measured values ​​of each secondary indicator in the multi-source basic data, and a fuzzy relation matrix is ​​constructed based on the membership degrees of all secondary indicators. The real-time weights of each level of indicators are dynamically calculated based on the mining process and the dynamic correlation coefficients of the aforementioned indicators. The membership degree of the criterion layer is calculated based on the real-time weights and the fuzzy relation matrix, and the entropy weight of the membership degree of the criterion layer is calculated. The weights of the criterion layer are corrected and the comprehensive membership degree is calculated. The corresponding early warning level and emergency response are obtained in the four-dimensional evaluation set according to the maximum value of the comprehensive membership degree.

[0006] As a further improvement of the present invention The static geological data includes: aquifer distribution, geological structure, old goaf register, and mechanical parameters of the coal seam roof and floor; The dynamic monitoring data includes: aquifer water pressure, working face water inflow, coal and rock deformation, water temperature and water level dynamics; The dynamic geophysical data includes: water-rich anomaly zone in front of the working face, development depth of water-conducting fracture zone, and fault water content; The hydrochemical data include: groundwater pH value, ion concentration, and isotope ratio.

[0007] As a further improvement to this invention, a hierarchical set of influencing factors, consisting of primary and secondary indicators, is constructed based on the water source conditions, water diversion channels, and dynamic factors of coal mine water inrush. This set includes: Water source conditions, water diversion channels, and dynamic factors are used as primary indicators. The secondary indicators corresponding to the water source conditions include the water-bearing capacity of the aquifer, the distribution of old-hole water, and the correlation with surface water. The secondary indicators corresponding to the water diversion channels include fault development, collapse columns, poor boreholes, and water-conducting fracture zones. The secondary indicators corresponding to the dynamic factors include changes in water pressure, changes in water inflow, and changes in hydrology.

[0008] As a further improvement of the present invention, the membership degree is calculated for the measured values ​​of each secondary indicator in the multi-source basic data, including: For the linear characteristics of the secondary indicators, a trapezoidal membership function is used to calculate the membership degree; for the nonlinear characteristics of the secondary indicators, a Gaussian membership function is used to calculate the membership degree; and for the coupled indicators of the secondary indicators, a coupled membership function is used in conjunction with the dynamic correlation coefficient to calculate the membership degree.

[0009] As a further improvement of the present invention, the real-time weights of each level of indicators are dynamically calculated based on the mining process and the dynamic correlation coefficients of the indicators; including: The analytic hierarchy process (AHP) is used to quantify the weights of each level of indicators to obtain static benchmark weights. Based on the dynamic correlation coefficient of the aforementioned indicators and the mining distance, a correction factor is calculated, and the static benchmark weight is corrected based on the correction factor to obtain the real-time weight.

[0010] As a further improvement of the present invention, the weights of each level of indicators are quantified using the analytic hierarchy process (AHP) to obtain static benchmark weights, including: The hierarchical structure consists of a target layer, a criterion layer, and an indicator layer. The target layer is the classification of flood inrush warnings, the criterion layer is the first-level indicators, including water source conditions, water diversion channels, and dynamic factors, and the indicator layer is all headphone indicators. Construct a weight judgment matrix, calculate the largest eigenvalue and weight vector, perform consistency checks, and obtain the total weight of the secondary indicators.

[0011] As a further improvement of the present invention, a correction factor is calculated based on the dynamic correlation coefficient of the indicators and the mining distance, and the static benchmark weight is corrected based on the correction factor to obtain the real-time weight, including: The formula for calculating the correction factor is: In the formula, k i-j This represents the correlation coefficient between secondary indicator i and secondary indicator j; ΔL represents the distance between the advancing distance of the mining face and the remaining distance to the danger point; L0 represents the initial distance; The formula for calculating real-time weights is: W i =W i0 ×α i / ∑（W i0 ×α i ） In the formula, W i0 Represents the static benchmark weights; α i This represents the correction factor.

[0012] As a further improvement of the present invention, the membership degree of the criterion layer is calculated based on the real-time weights and the fuzzy relation matrix, and the entropy weight of the membership degree of the criterion layer is calculated. The criterion layer weights are then corrected and the comprehensive membership degree is calculated, including: The formula for calculating the membership degree of the criterion layer is as follows: In the formula, W k R represents the real-time weight of the secondary indicator. k This represents the corresponding fuzzy submatrix; The formula for the weight of the correction criterion layer is: W u '=W u ×E k / ∑(W u ×E k ) In the formula, W u E represents the weight of the criterion layer. k Entropy weights representing the membership degree of the criterion layer; The formula for calculating the overall membership degree is: In the formula, W u ' indicates the weight of the modified criterion layer, and B1, B2, and B3 represent the membership degree of the criterion layer.

[0013] As a further improvement of the present invention, the corresponding early warning level and emergency response are obtained from the four-dimensional evaluation set based on the maximum value in the comprehensive membership degree; including: Based on the early warning level and the implementation monitoring data, the prevention and control engineering parameter library is called to output matching parameters, including grouting parameters, drainage parameters and monitoring parameters; Generate and publish an early warning report, which includes the level, probability of water inrush, emergency parameters, risk area, and three-dimensional coordinates; push the early warning report to the downhole display screen, the mobile terminal of the management personnel, and the dispatch center.

[0014] As a further improvement of the present invention, a closed-loop iterative mechanism of early warning-handling-feedback is established to track the effect of early warning, correct the model, and iterate the model.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention proposes a coal mine water inrush early warning and classification method. Based on the whole chain approach of "multi-source data fusion → dynamic weight allocation → intelligent fuzzification → emergency linkage", it integrates AHP - improved fuzzy evaluation and Internet of Things (IoT) technology to construct an integrated early warning method of "precise identification - scientific classification - efficient handling". This fills the technical gap in refined early warning of water inrush under complex conditions and realizes "precise identification, dynamic classification and efficient response" of water inrush risk.

[0016] This invention improves the accuracy of coal mine water anomaly identification and reduces the false judgment rate by integrating multi-dimensional data sources; enhances early warning adaptability by updating weights in real time with the mining process; shortens on-site response time and improves the timeliness of water control by matching precise prevention and control parameters; solves the quantification problem of nonlinear indicators by using a combination of various membership functions; and shortens the adaptation time for new mines by using an iterative model based on actual cases. Attached Figure Description

[0017] Figure 1 This is a flowchart of a coal mine water inrush early warning and classification method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the hierarchical structure of factors affecting water inrush in coal mines, as disclosed in one embodiment of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] The present invention will now be described in further detail with reference to the accompanying drawings: like Figure 1 As shown, the present invention provides a coal mine water inrush early warning and classification method, which integrates multi-source data fusion technology, dynamic AHP weighting, improved fuzzy evaluation, and emergency parameter matching technology, including: S1. Real-time acquisition of multi-source basic data in coal mining areas, including static geological data, dynamic detection data, dynamic geophysical data, and hydrochemical data; in, Static geological data includes: aquifer distribution, geological structure, old goaf register, and mechanical parameters of the coal seam roof and floor; Dynamic monitoring data includes: aquifer water pressure, working face water inflow, coal and rock deformation, water temperature and water level dynamics; Dynamic geophysical data includes: water-rich anomaly zone ahead of the working face, depth of water-conducting fracture zone development, and fault water content; The water chemistry data includes: groundwater pH, ion concentration, and isotope ratios.

[0020] Furthermore, The distribution of water layers includes thickness, burial depth, and lithology; geological structures include fault / collapse column orientation and fracture zone width; old goaf records include water accumulation range and water level elevation; and the mechanical parameters of the coal seam roof and floor include elastic modulus and impermeability. Aquifer water pressure includes static / dynamic pressure, and coal and rock deformation includes the approach of the roof and floor. The location and size of the anomalous water-rich area in front of the work site; Furthermore, Static mechanism data acquisition methods include: geological borehole columnar sections, 3D seismic exploration reports, and laboratory rock sample tests; Dynamic monitoring data acquisition equipment includes: water pressure sensor, electromagnetic flow sensor, fiber optic displacement gauge, and temperature sensor; Dynamic geophysical data acquisition equipment includes: transient electromagnetic instruments, ground-penetrating radar, borehole imaging instruments, etc. The method for collecting water chemical data is as follows: collect water samples and perform tests; Specifically, The collected data undergoes data standardization preprocessing, and then an integrated database consisting of a three-dimensional geological model, a dynamic monitoring time series database, a geophysical anomaly database, and a hydrochemical characteristic database is constructed to support real-time data retrieval and related queries.

[0021] S2. Based on the water source conditions, water diversion channels, and dynamic factors of coal mine water inrush, a hierarchical set of primary and secondary indicators is constructed, as shown in Table 1. The definition, quantification method, and data source of each indicator are clarified to ensure that they are directly related to the risk of water inrush. Dynamic correlation coefficients of the indicators are also introduced. in, Water source conditions, water diversion channels, and dynamic factors are used as primary indicators; Secondary indicators corresponding to water source conditions include aquifer water abundance, distribution of old dead water, and correlation with surface water; secondary indicators corresponding to water conduction channels include fault development, collapse columns, poor boreholes, and water-conducting fracture zones; secondary indicators corresponding to dynamic factors include changes in water pressure, changes in inflow, and changes in hydrology.

[0022] Table 1 S3. Define the four-dimensional evaluation set of early warning level, symbol, evaluation criteria and emergency response for coal mine water hazards, as shown in Table 2. in, The four-dimensional evaluation set can be represented as V = {V1, V2, V3, V4}; Table 2 S4. Calculate the membership degree for the measured values ​​of each secondary indicator in the multi-source basic data, and construct a fuzzy relation matrix based on the membership degrees of all secondary indicators. in, To address the nonlinear characteristics of different indicators, a combined model of "trapezoidal membership function + Gaussian membership function + coupled membership function" is used to calculate the membership degree r.ij This addresses the issue of insufficient resolution inherent in traditional single-function methods. This includes: using a trapezoidal membership function to calculate membership degree for linear characteristics in secondary indicators; using a Gaussian membership function to calculate membership degree for nonlinear characteristics in secondary indicators; and using a coupled membership function to calculate membership degree for coupled indicators in secondary indicators in combination with dynamic correlation coefficients.

[0023] Specifically, (1) Trapezoidal membership function: applicable to linear indices (such as the distance between old and new water levels) ① Positive correlation type: ② Negative correlation type: In the formula: ui is the measured value of the secondary indicator; aj, bj, and cj are the boundaries of the interval to which the corresponding warning level Vj belongs. (2) Gaussian membership function: applicable to strongly nonlinear indices (such as water pressure, flow rate), formula: In the formula: bj is the optimal value of the indicator for the warning level Vj; σj is the standard deviation (determined based on historical data statistics); (3) Coupling membership function: Applicable to coupling indices such as "water pressure-inflow" and "fracture zone-fault", formula: In the formula: u i u k For coupling index; k i-k This is the dynamic correlation coefficient.

[0024] (4) Construction of fuzzy matrix: The membership degrees of 10 secondary indicators constitute a 10×4 improved fuzzy relation matrix R, and the matrix elements include the membership degree of single indicators and the membership degree of coupled indicators.

[0025] S5. Calculate the real-time weights of each level of indicators based on the mining process and the dynamic correlation coefficient of the indicators. in, The analytic hierarchy process (AHP) is used to quantify the weights of each level of indicators to obtain static benchmark weights. Based on the dynamic correlation coefficient of the indicators and the mining distance, a correction factor is calculated, and the static benchmark weight is corrected based on the correction factor to obtain the real-time weight.

[0026] Furthermore, The analytic hierarchy process (AHP) is used to quantify the weights of each level of indicators, resulting in static benchmark weights, including: Hierarchical structure such as Figure 2As shown, it includes a target layer, a criterion layer, and an indicator layer. The target layer is the early warning classification of water inrush, the criterion layer is the first-level indicator, including water source conditions, water diversion channels and dynamic factors, and the indicator layer is all headphone indicators. Construct a weight judgment matrix, calculate the largest eigenvalue and weight vector, perform consistency checks, and obtain the total weight of the secondary indicators.

[0027] Based on the dynamic correlation coefficient of the indicators and the mining distance, a correction factor is calculated. The static baseline weights are then corrected based on this correction factor to obtain the real-time weights, including: The formula for calculating the correction factor is: In the formula, k i-j This represents the correlation coefficient between secondary indicator i and secondary indicator j; ΔL represents the distance between the advancing distance of the mining face and the remaining distance to the danger point; L0 represents the initial distance; The formula for calculating real-time weights is: W i =W i0 ×α i / ∑（W i0 ×α i ） In the formula, W i0 Represents the static benchmark weights; α i This represents the correction factor.

[0028] S6. Calculate the membership degree of the criterion layer based on the real-time weight and fuzzy relation matrix, and calculate the entropy weight of the membership degree of the criterion layer. Correct the weight of the criterion layer and calculate the comprehensive membership degree. Obtain the corresponding early warning level and emergency response in the four-dimensional evaluation set according to the maximum value in the comprehensive membership degree.

[0029] in, The membership degree of the criterion layer is calculated based on real-time weights and fuzzy relation matrices, and the entropy weight of the membership degree of the criterion layer is calculated. The criterion layer weights are then corrected and the comprehensive membership degree is calculated, including: The formula for calculating the membership degree of the criterion layer is: In the formula, W k R represents the real-time weight of the secondary indicator. k This represents the corresponding fuzzy submatrix; The formula for adjusting the criterion layer weights is as follows: Wu '=W u ×E k / ∑(W u ×E k ) In the formula, W u E represents the weight of the criterion layer. k Entropy weights representing the membership degree of the criterion layer; The formula for calculating the overall membership degree is: In the formula, W u ' indicates the weight of the modified criterion layer, and B1, B2, and B3 represent the membership degree of the criterion layer.

[0030] Furthermore, the corresponding early warning level and emergency response are obtained from the four-dimensional evaluation set based on the maximum value in the comprehensive membership degree; including: Based on the early warning level and the implementation monitoring data, the prevention and control engineering parameter library is called to output matching parameters, including grouting parameters, drainage parameters and monitoring parameters; Generate and publish early warning reports, which include the level, probability of water inrush, emergency parameters, risk area, and three-dimensional coordinates; push the early warning reports to downhole display screens, mobile terminals of management personnel, and dispatch centers.

[0031] The present invention also includes establishing a closed-loop iterative mechanism of early warning-handling-feedback, tracking the effect of early warning, correcting the model, and iterating the model.

[0032] Example 1: like Figure 1 , 2 As shown, the specific implementation process of the present invention for early warning and classification of coal mine water inrush hazards includes: Step 1: Multi-source basic data acquisition and preprocessing Based on a four-dimensional data source of "static geology + dynamic monitoring + geophysical exploration + hydrochemical analysis", a basic database of water hazards is constructed and preprocessed. (1) Static geological data: Contents: Aquifer distribution (thickness, burial depth, lithology), geological structure (fault / collapse column orientation, fracture zone width), old goaf inventory (water accumulation range, water level elevation), and mechanical parameters of the coal seam roof and floor (elastic modulus, impermeability). Data collection methods: geological borehole columnar section, 3D seismic exploration report, and laboratory rock sample testing; (2) Dynamic monitoring data: Contents: Aquifer water pressure (static / dynamic pressure), working face water inflow, coal and rock deformation (roof and floor approach), water temperature, and water level dynamics; Data acquisition equipment: water pressure sensor, electromagnetic flow sensor, fiber optic displacement gauge, temperature sensor; (3) Dynamic geophysical data: Contents: Water-rich anomaly zone ahead of the working face (location, scale), depth of water-conducting fracture zone development, and water-bearing properties of faults; Data acquisition equipment: transient electromagnetic instrument, ground-penetrating radar, borehole imaging instrument, etc.; (4) Hydrochemical data: Contents: Groundwater pH value, ion concentration, and isotope ratio; Collection method: Collect water samples and test them; Data fusion output: After data standardization and preprocessing, an integrated database is constructed consisting of "3D geological model + dynamic monitoring time series database + geophysical anomaly database + hydrochemical characteristic database", which supports real-time data access and related queries.

[0033] Step 2: Construct a system of influencing factors based on "three-level indicators and dynamic correlation". Based on the three-element disaster logic of coal mine water inrush – "water source-channel-dynamic factors" – a hierarchical influencing factor set U={U1,U2,U3} of "first-level indicators-second-level indicators" is constructed, as shown in Table 1. The definitions, quantification methods, and data sources of each indicator are clearly defined to ensure direct correlation with water inrush risk. A "dynamic correlation coefficient of indicators" is introduced to reflect the coupling effect between factors. The function of the dynamic correlation coefficient is to automatically adjust the contribution of the secondary indicators when the value of the coupling indicator changes.

[0034] Step 3: Determine the grading and evaluation set for flood inrush early warning. Based on the severity of sudden flooding, a four-dimensional assessment set V = {V1, V2, V3, V4} is defined, consisting of "risk level - assessment criteria - emergency response," corresponding to four warning levels, as shown in Table 2. Step 4: Intelligent Fuzzification of Indicators (Constructing an Improved Fuzzy Relation Matrix R) To address the nonlinear characteristics of different indicators, a combined model of "trapezoidal membership function + Gaussian membership function + coupled membership function" is used to calculate the membership degree r. ij This addresses the issue of insufficient resolution inherent in traditional single-function methods. (1) Trapezoidal membership function: applicable to linear indices (such as the distance between old and new water levels) ① Positive correlation type: ② Negative correlation type: In the formula: u i The measured value of the secondary indicator; a j b j c j The corresponding early warning level V for this indicatorj The boundary of the membership interval.

[0035] (2) Gaussian membership function: applicable to strongly nonlinear indices (such as water pressure, flow rate), formula: In the formula: b j Warning level V j The optimal value of the index; σ j The standard deviation is (determined based on historical data statistics). (3) Coupling membership function: Applicable to coupling indices such as "water pressure-inflow" and "fracture zone-fault", formula: In the formula: u i u k For coupling index; k i-k This is the dynamic correlation coefficient.

[0036] Fuzzy matrix construction: The membership degrees of 10 secondary indicators constitute a 10×4 improved fuzzy relation matrix R, and the matrix elements include the membership degree of single indicators and the membership degree of coupled indicators.

[0037] Step 5: Dynamic AHP weight calculation (introducing a time-series update mechanism) The real-time weights are calculated using a combination of static baseline weights and dynamic correction coefficients to ensure that the weights dynamically adapt to the mining process. (1) Calculation of static benchmark weights: The Analytic Hierarchy Process (AHP) is used to quantify the weights of indicators at each level, ensuring a scientific and reasonable weight allocation. Specific steps include: ① Hierarchical structure: Target layer (flood disaster early warning classification) → Criterion layer (U1 water source, U2 channel, U3 dynamic factors) → Indicator layer (10 secondary indicators); ② Judgment matrix construction: Experts used the "1~9 scale method + Delphi method" to construct the judgment matrix and calculated the maximum eigenvalue λmax and the weight vector; ③ Consistency test: CR=CI / RI, CI=(λmax-n) / (n-1), CR<0.1 is required to obtain the total weight W of the secondary indicators.

[0038] Table 3 (2) Dynamic weight adjustment: Correction factor: The correction factor α is calculated based on the "dynamic correlation coefficient of the indicator + mining distance". i : In the formula: ki-j is the correlation coefficient between index i and core index j; ΔL is the working face advance distance and the remaining distance to dangerous points (such as faults); L0 is the initial distance; Real-time weight: W i = W i0 ×α i / ∑(W i0 ×α i ) (W i0 (Static benchmark weights).

[0039] Weighted time series update: Every 5m advance or when a new anomaly is discovered during geophysical exploration, the weights are automatically recalculated, and a weight change curve is generated.

[0040] Step 6: Improve fuzzy comprehensive evaluation (introduce entropy weight-weighted average combination operator) A three-level fuzzy evaluation method is adopted, which combines "entropy weight correction" and "weighted average method" to improve the objectivity of the results. (1) First-level evaluation (second-level indicators → criteria layer): Criterion layer membership degree: (W) k R represents the real-time weight of the secondary indicator. k (corresponding to the fuzzy submatrix); (2) Second-level evaluation (criteria level → target level): Introducing entropy weight correction: calculating the entropy weight E of the membership degree of the criterion layer. k Adjust the weights of the criterion layer: W u '=W u ×E k / ∑(W u ×E k ); Overall membership degree: .

[0041] (3) Early warning level determination: Based on the principle of maximum membership, the warning level corresponding to the maximum value in B is selected.

[0042] Step 7: Emergency Response Parameter Matching (1) Emergency parameter matching: Based on the early warning level and real-time monitoring data, the "Prevention and Control Engineering Parameter Library" is invoked to output matching parameters: Grouting parameters: grouting pressure, grouting volume, and grout mix ratio (calculated based on water pressure and fracture zone height); Drainage parameters: Number of new drainage pumps and diameter of drainage pipes required (calculated based on inflow and existing drainage capacity); Monitoring parameters: frequency of encrypted monitoring, location of newly added monitoring points.

[0043] (2) Issuance of early warning information: The system generates an early warning report containing "level, probability of water inrush, emergency parameters, and three-dimensional coordinates of the risk area" and pushes it to the downhole display screen, the mobile terminal of the management personnel, and the dispatch center.

[0044] Step 8: Early Warning Effect Feedback and Model Iteration Establish a closed-loop iterative mechanism of "early warning-response-feedback": (1) Effect tracking: Record whether water inrush occurs after the warning and the effectiveness of the treatment project (e.g., the water inrush volume decreases after grouting); (2) Model correction: The membership function parameters and weight correction factors are corrected every 3 months based on actual data; (3) Model iteration: Construct an "Engineering Case Library - Early Warning Model Parameter Library" to achieve rapid model adaptation for new mines and working faces.

[0045] Advantages of this invention: This invention proposes a coal mine water inrush early warning and classification method. Based on the whole chain approach of "multi-source data fusion → dynamic weight allocation → intelligent fuzzification → emergency linkage", it integrates AHP - improved fuzzy evaluation and Internet of Things (IoT) technology to construct an integrated early warning method of "precise identification - scientific classification - efficient handling". This fills the technical gap in refined early warning of water inrush under complex conditions and realizes "precise identification, dynamic classification and efficient response" of water inrush risk.

[0046] This invention improves the accuracy of coal mine water anomaly identification and reduces the false judgment rate by integrating multi-dimensional data sources; enhances early warning adaptability by updating weights in real time with the mining process; shortens on-site response time and improves the timeliness of water control by matching precise prevention and control parameters; solves the quantification problem of nonlinear indicators by using a combination of various membership functions; and shortens the adaptation time for new mines by using an iterative model based on actual cases.

[0047] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for early warning and classification of water inrush in coal mines, characterized in that, include: Real-time acquisition of multi-source basic data in coal mining areas, including static geological data, dynamic detection data, dynamic geophysical data, and hydrochemical data; Based on the water source conditions, water diversion channels, and dynamic factors of coal mine water inrush, a hierarchical set of influencing factors consisting of primary and secondary indicators is constructed. The definitions, quantification methods, and data sources of each indicator are clarified, and dynamic correlation coefficients of the indicators are introduced. Define a four-dimensional evaluation set for early warning levels, symbols, evaluation criteria, and emergency response to coal mine water hazards; The membership degree is calculated for the measured values ​​of each secondary indicator in the multi-source basic data, and a fuzzy relation matrix is ​​constructed based on the membership degrees of all secondary indicators. The real-time weights of each level of indicators are dynamically calculated based on the mining process and the dynamic correlation coefficients of the aforementioned indicators. The membership degree of the criterion layer is calculated based on the real-time weights and the fuzzy relation matrix, and the entropy weight of the membership degree of the criterion layer is calculated. The weights of the criterion layer are corrected and the comprehensive membership degree is calculated. The corresponding early warning level and emergency response are obtained in the four-dimensional evaluation set according to the maximum value of the comprehensive membership degree.

2. The coal mine water hazard inrush early warning and classification method according to claim 1, characterized in that: The static geological data includes: aquifer distribution, geological structure, old goaf register, and mechanical parameters of the coal seam roof and floor; The dynamic monitoring data includes: aquifer water pressure, working face water inflow, coal and rock deformation, water temperature and water level dynamics; The dynamic geophysical data includes: water-rich anomaly zone in front of the working face, development depth of water-conducting fracture zone, and fault water content; The hydrochemical data include: groundwater pH value, ion concentration, and isotope ratio.

3. The coal mine water hazard inrush early warning and classification method according to claim 1, characterized in that: Based on the water source conditions, water diversion channels, and dynamic factors of coal mine water inrush, a hierarchical set of influencing factors consisting of primary and secondary indicators is constructed, including: Water source conditions, water diversion channels, and dynamic factors are used as primary indicators. The secondary indicators corresponding to the water source conditions include the water-bearing capacity of the aquifer, the distribution of old-hole water, and the correlation with surface water. The secondary indicators corresponding to the water diversion channels include fault development, collapse columns, poor boreholes, and water-conducting fracture zones. The secondary indicators corresponding to the dynamic factors include changes in water pressure, changes in water inflow, and changes in hydrology.

4. The coal mine water hazard inrush early warning and classification method according to claim 1, characterized in that: The membership degree is calculated for the measured values ​​of each secondary indicator in the multi-source basic data, including: For the linear characteristics of the secondary indicators, a trapezoidal membership function is used to calculate the membership degree; for the nonlinear characteristics of the secondary indicators, a Gaussian membership function is used to calculate the membership degree; and for the coupled indicators of the secondary indicators, a coupled membership function is used in conjunction with the dynamic correlation coefficient to calculate the membership degree.

5. The coal mine water hazard inrush early warning and classification method according to claim 1, characterized in that: The real-time weights of each level of indicator are dynamically calculated based on the mining process and the dynamic correlation coefficients of the aforementioned indicators; including: The analytic hierarchy process (AHP) is used to quantify the weights of each level of indicators to obtain static benchmark weights. Based on the dynamic correlation coefficient of the aforementioned indicators and the mining distance, a correction factor is calculated, and the static benchmark weight is corrected based on the correction factor to obtain the real-time weight.

6. The coal mine water hazard inrush early warning and classification method according to claim 5, characterized in that: The analytic hierarchy process (AHP) is used to quantify the weights of each level of indicators, resulting in static benchmark weights, including: The hierarchical structure consists of a target layer, a criterion layer, and an indicator layer. The target layer is the classification of flood inrush warnings, the criterion layer is the first-level indicators, including water source conditions, water diversion channels, and dynamic factors, and the indicator layer is all headphone indicators. Construct a weight judgment matrix, calculate the largest eigenvalue and weight vector, perform consistency checks, and obtain the total weight of the secondary indicators.

7. The coal mine water hazard inrush early warning and classification method according to claim 5, characterized in that: Based on the dynamic correlation coefficient of the aforementioned indicators and the mining distance, a correction factor is calculated. The static baseline weights are then corrected based on this correction factor to obtain real-time weights, including: The formula for calculating the correction factor is: In the formula, k i-j This represents the correlation coefficient between secondary indicator i and secondary indicator j; ΔL represents the distance between the advancing distance of the mining face and the remaining distance to the danger point; L0 represents the initial distance; The formula for calculating real-time weights is: W i =W i0 ×α i / ∑（W i0 ×α i ） In the formula, W i0 Represents the static benchmark weight; α i This represents the correction factor.

8. The coal mine water hazard inrush early warning and classification method according to claim 1, characterized in that: The membership degree of the criterion layer is calculated based on the real-time weights and the fuzzy relation matrix, and the entropy weight of the membership degree of the criterion layer is calculated. The criterion layer weights are then corrected, and the comprehensive membership degree is calculated, including: The formula for calculating the membership degree of the criterion layer is as follows: In the formula, W k R represents the real-time weight of the secondary indicator. k This represents the corresponding fuzzy submatrix; The formula for the weight of the correction criterion layer is: IN u '=W u ×E k / ∑(W u ×E k ) In the formula, W u E represents the weight of the criterion layer. k Entropy weight representing the membership degree of the criterion layer; The formula for calculating the overall membership degree is: In the formula, W u ' indicates the weight of the modified criterion layer, and B1, B2, and B3 represent the membership degree of the criterion layer.

9. The coal mine water hazard inrush early warning and classification method according to claim 1, characterized in that: Based on the maximum value in the comprehensive membership degree, the corresponding early warning level and emergency response are obtained from the four-dimensional evaluation set; including: Based on the early warning level and the implementation monitoring data, the prevention and control engineering parameter library is called to output matching parameters, including grouting parameters, drainage parameters and monitoring parameters; Generate and publish an early warning report, which includes the level, probability of water inrush, emergency parameters, risk area, and three-dimensional coordinates; push the early warning report to the downhole display screen, the mobile terminal of the management personnel, and the dispatch center.

10. The coal mine water hazard inrush early warning and classification method according to claim 1, characterized in that: Establish a closed-loop iterative mechanism of early warning, response, and feedback to track the effectiveness of early warning, revise the model, and iterate the model.

Images