A Comprehensive Risk Assessment Method for Geological Hazards in Coal Mine Power Supply Lines Based on Fuzzy Logic and Multi-Criterion Decision Making

Abstract

This invention belongs to the field of coal mine power safety and intelligent risk management technology, and discloses a comprehensive risk rating method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making. The method first collects multi-dimensional indicator data and completes hierarchical downgrading processing to construct a four-dimensional, three-layer geological disaster risk indicator system; an initial risk score is obtained through hierarchical fuzzy reasoning, and a geological disaster recoverability time correction function is introduced to correct the risk score; the score is calibrated through the fusion of subjective and objective weights and a conflict self-checking mechanism, and the entire process of risk rating update is triggered based on a preset dynamic threshold. This invention solves the problems of premature resolution of fuzzy information, lack of information conflict self-checking, and failure to consider recovery time in existing technologies. It retains the advantages of fuzzy reasoning uncertainty, avoids rule combination explosion, can proactively identify weight conflicts and circuit breakers for error correction, improves the robustness of risk assessment, and accurately matches the needs of coal mine power supply line repair resource allocation and safety management.

Description

Technical Field

[0001] This invention belongs to the field of coal mine power safety and intelligent risk management technology, specifically involving a comprehensive risk rating method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making. Background Technology

[0002] Power supply lines in coal mines are critical infrastructure for mine production. Interruptions due to geological disasters (such as landslides, goaf subsidence, and debris flows) will directly lead to the failure of ventilation and drainage systems. Because of the risk of gas accumulation underground, a power outage necessitating a ventilation shutdown could have catastrophic consequences. The geological hazard risk assessment of coal mine power supply lines faces the following industry-specific technical challenges:

[0003] 1) The contradiction between the timeliness of risks and the limited nature of resources: Mine repair resources are limited, and priority should be given to preventing potential hazards that are not only prone to occur and have serious consequences, but also have long recovery times. Traditional risk models only focus on the probability and severity of occurrence and cannot identify "high-impact sections that are difficult to repair," leading to a mismatch between resource allocation strategies and on-site safety needs.

[0004] 2) The contradiction between multi-source data fusion and ambiguity preservation: The assessment requires the fusion of multi-source data, such as geological (static), meteorological (dynamic), and inspection (subjective). If ambiguous indicators are clarified in the early stages of fusion (e.g., through simple weighted summation), the uncertain information in the data will be lost, weakening the robustness of subsequent reasoning.

[0005] 3) The contradiction between expert experience and dynamic objective data: When sensor data is distorted due to faults, or expert experience becomes outdated due to changes in geological conditions, the weighted evaluation results of the two will conflict drastically. Existing methods lack identification and self-checking mechanisms for such conflicts, which may lead to unreliable evaluation conclusions.

[0006] Existing technical solutions and their limitations:

[0007] Currently, there are studies applying fuzzy logic and multi-criteria decision-making to engineering risk assessment. A typical implementation involves first determining the weights of indicators using the analytic hierarchy process (AHP), and then using fuzzy comprehensive evaluation to classify the risks into levels. However, the above approach still has the following limitations when dealing with geological disaster assessments of coal mine power supply lines:

[0008] 1) Ambiguous information is prematurely eliminated during the aggregation process: Existing methods usually normalize the values ​​of the indicators before weighted aggregation, which causes the ambiguity of the indicators to be lost before entering the core of reasoning.

[0009] 2) Lack of self-checking and circuit breaker mechanisms for information conflicts: When subjective weights and objective weights are in serious conflict, existing methods still force them to merge, which masks data quality or model applicability issues.

[0010] 3) Risk model does not reflect recovery time: For power supply lines, for the same degree of damage, the longer the recovery time, the greater the accumulation of downhole risks. Existing models fail to quantify this difference. Summary of the Invention

[0011] This invention aims to provide a comprehensive risk assessment method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making.

[0012] To achieve the above objectives, the present invention adopts the following technical solution:

[0013] A comprehensive risk assessment method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making, the method includes the following steps:

[0014] Step S1: Collect multi-dimensional indicator data of potential geological hazards in coal mine power supply lines, classify data sources according to confidence level, remove missing indicators and redistribute weights proportionally to obtain a standardized indicator dataset.

[0015] Step S2: Based on geological background, topography, meteorology and hydrology, and engineering management, construct a first-level, second-level, and third-level geological disaster risk indicator system, and match the standardized indicator dataset to the corresponding third-level indicators;

[0016] Step S3: The matched three-level indicators are fuzzified using the adaptive membership function. After aggregation by primary fuzzy inference, three scores are obtained: probability of occurrence, severity, and detectability. The three scores are then input into the advanced fuzzy inference engine to output the initial risk score.

[0017] Step S4: Based on the initial risk score, introduce the geological disaster recovery time factor, calculate the correction coefficient through the risk correction function, and obtain the risk score corrected for recovery time.

[0018] Preferably, the method further includes the following steps:

[0019] Step S5: Calculate the subjective and objective weights of the indicators, convert the weights into the basic probability distribution of evidence theory, measure the degree of conflict between subjective and objective weights and perform fusion or circuit breaking processing, and calibrate the risk score after recovery time correction using the comprehensive weights.

[0020] Step S6: Based on the preset configurable threshold, monitor the dynamic parameters. When the preset triggering conditions are met, use the calibrated risk score as a benchmark and re-execute steps S1 to S5 to complete the dynamic update of the risk rating throughout the entire process.

[0021] Preferably, in step S1, the data source is divided into three levels of confidence based on measured data, historical statistical values, and expert experience assignments, and the weight of missing indicators is redistributed according to the proportion of remaining indicators belonging to the same higher-level indicators.

[0022] Preferably, the membership function in step S3 includes three types: trapezoidal, triangular, and S-shaped, which are respectively adapted to the three-level index with standardized threshold, design optimal value, and nonlinear gradual change characteristics.

[0023] Preferably, in step S3, the primary fuzzy inference group and aggregate the three-level indicators into single-value scores, and the advanced fuzzy inference uses the Mandani fuzzy inference mechanism and the centroid method to defuzzify and output the initial risk score.

[0024] Preferably, the rule size of the primary fuzzy inference in step S3 is controlled, and the rule base of the advanced fuzzy inference is composed of low / medium / high linguistic values ​​of occurrence probability, severity, and detectability.

[0025] Preferably, the calculation formulas for the risk correction function and the risk score adjusted for recovery time in step S4 are as follows:

[0026] ;

[0027] ;

[0028] In the formula, Scoring the recovery time factor for geological disasters; This is the maximum score. This is the adjustment coefficient; For the initial risk score; This is a risk score adjusted for recovery time.

[0029] Preferably, the subjective weights in step S5 are calculated using the triangular fuzzy number hierarchical analysis method, and the objective weights are calculated using the information entropy method.

[0030] Preferably, in step S5, the Jussellmey evidence distance metric is used to weigh the conflict degree. When the conflict degree exceeds a preset threshold, circuit breaking is performed and subjective weighting is applied.

[0031] Preferably, when converting the weights into basic probability allocation in step S5, a preset reliability coefficient is used for discounting.

[0032] The present invention provides a comprehensive risk assessment method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making, which achieves several technical advantages:

[0033] 1. By using a hierarchical fuzzy inference architecture, 36 high-dimensional indicators are effectively mapped to 3 low-dimensional risk variables. This not only retains the advantages of fuzzy inference in handling uncertainty, but also avoids the rule combinatorial explosion problem, making the inference system maintainable in engineering.

[0034] 2. By introducing a geological disaster recovery time correction function, the risk rating results can reflect the cumulative impact of recovery time, which is more in line with the emergency repair time requirements of mine power supply lifeline projects and is conducive to optimizing the priority allocation of emergency repair resources.

[0035] 3. Through conflict self-checking and circuit breaker mechanisms, it can proactively identify serious conflicts between subjective and objective information and interrupt the fusion process in a timely manner, preventing erroneous ratings caused by sensor failures or expert misjudgments, and improving the robustness of the evaluation results. Attached Figure Description

[0036] Figure 1 The flowchart shows steps S1-S4 of the comprehensive risk rating method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making of the present invention.

[0037] Figure 2 This is a schematic diagram of the hierarchical fuzzy inference architecture of the present invention.

[0038] Figure 3 This is a schematic diagram of the TTR risk continuous correction function relationship of the present invention.

[0039] Figure 4 This is a flowchart of steps S1-S6 of the comprehensive risk rating method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making of the present invention. Detailed Implementation

[0040] The following detailed implementation of the present invention, based on fuzzy logic and multi-criteria decision-making, provides a comprehensive risk assessment method for geological disaster hazards in coal mine power supply lines. These embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0041] The essential features that distinguish this invention from the prior art are as follows:

[0042] 1) Hierarchical fuzzy inference architecture: Through a hierarchical structure of indicator grouping, primary fuzzy inference, defuzzification aggregation, and advanced fuzzy inference, high-dimensional indicators are mapped to low-dimensional risk variables, effectively avoiding the rule combinatorial explosion problem while retaining fuzzy information.

[0043] 2) Risk correction function for continuous geological disaster recovery time: In view of the industry-specific nature of mines where power outages mean ventilation outages, the post-disaster recovery capacity is explicitly introduced into the risk model in the form of a continuous function, so that the assessment results can reflect the cumulative impact of recovery time.

[0044] 3) Weight fusion conflict self-check and circuit breaker mechanism: The evidence distance is used to measure the degree of conflict between subjective and objective weight vectors. When the conflict exceeds the preset threshold, the fusion is actively interrupted and the system is rolled back to a reliable weight source to avoid the amplification of errors in the multi-source information fusion process.

[0045] Example 1: Implementation of a comprehensive risk rating method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making.

[0046] Combined with appendix Figures 1-3 As shown, this invention provides a comprehensive risk rating method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making.

[0047] like Figure 1 As shown, Figure 1 This is a flowchart of steps S1-S4 of the comprehensive risk rating method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making in Embodiment 1 of the present invention.

[0048] Step S1: Data source grading and dynamic degradation mechanism.

[0049] Step S1 involves collecting multi-dimensional indicator data of potential geological hazards along coal mine power supply lines, classifying data sources according to confidence levels, removing missing indicators and redistributing weights proportionally to obtain a standardized indicator dataset.

[0050] In step S1, the data source is divided into three levels of confidence based on measured data, historical statistical values, and expert experience. The weights of missing indicators are redistributed according to the proportion of remaining indicators belonging to the same higher-level indicator. The specific process is as follows:

[0051] To ensure project availability under conditions of incomplete data, a data source classification and dynamic degradation mechanism is established, as shown in Table 1, which contains the data source classification and degradation rules.

[0052] Table 1

[0053]

[0054] The confidence weight β is derived by analyzing the consistency rate between various data sources and actual situations in geological disaster events in the mining area over the past 5 years. When data for a certain tertiary indicator is completely missing and there is no substitute value of grade B or C, that indicator is removed from the current assessment task, and its weight is redistributed proportionally to other tertiary indicators belonging to the same higher-level indicator according to the following formula:

[0055] ;

[0056] in, As an indicator The original weights, The weight of the removed indicator, For the redistributed weights, The original weight of the v-th indicator that was not removed. To retain the original total weight of the indicators.

[0057] Step S2: Construction of the indicator system.

[0058] Step S2 constructs a primary, secondary, and tertiary geological disaster risk indicator system based on geological background, topography, meteorology, hydrology, and engineering management, and matches the standardized indicator dataset to the corresponding tertiary indicators. The specific process is as follows:

[0059] Based on the "Geological Hazard Risk Assessment Standard" (DZ / T 0286-2015), the "Coal Mine Safety Regulations," and the "Detailed Rules for Coal Mine Water Prevention and Control," an assessment system comprising 4 primary indicators, 12 secondary indicators, and 36 tertiary indicators was constructed. Of these, 34 indicators are derived from the aforementioned standards, while 2 (traffic accessibility index and backup line switching capability) are supplementary indicators for power supply line scenarios. Table 2 shows the list of tertiary indicators.

[0060] Table 2

[0061]

[0062]

[0063] Step S3, hierarchical fuzzy reasoning.

[0064] In step S3, the matched three-level indicators are fuzzified using the appropriate membership function. After primary fuzzy inference aggregation, three scores are obtained: probability of occurrence, severity, and detectability. The three scores are then input into the advanced fuzzy inference engine to output the initial risk score.

[0065] The membership functions mentioned in step S3 include three types: trapezoidal, triangular, and S-shaped, which are respectively adapted to three-level indices with standardized threshold, design optimal value, and nonlinear gradual change characteristics.

[0066] In step S3, the primary fuzzy inference group and aggregate the three-level indicators into single-value scores, while the advanced fuzzy inference uses the Mandani fuzzy inference mechanism and the centroid method to defuzzify and output the initial risk score.

[0067] In step S3, the rule size for primary fuzzy inference is controlled, while the rule base for advanced fuzzy inference consists of low / medium / high linguistic value combinations of occurrence probability, severity, and detectability. The specific process is as follows:

[0068] (1) Indicator fuzziness

[0069] For each tertiary indicator, fuzzification is performed using trapezoidal membership functions, triangular membership functions, or S-shaped membership functions, depending on its data type and engineering characteristics.

[0070] Trapezoidal membership functions are suitable for indicators with clearly defined threshold values, such as slope, and their expression is:

[0071] ;

[0072] in, It is a trapezoidal membership function. The input variables to be fuzzed (such as the slope angle). , , , This refers to the shape parameter of the membership function, specifically the angular shape parameter of the trapezoidal membership function, set according to the "Code for Geological Hazard Risk Assessment" (DZ / T 0286-2015). Taking the language value of a gentle slope as an example, the parameter is set as follows: , , , That is, the slope angle is to When the slope is between 1 and 1, the membership degree of the slope is 1; when the slope angle exceeds 1, the membership degree of the slope is 1. The membership degree of the gentle slope then decreased to 0.

[0073] The triangular membership function is applicable to indices with optimal design values ​​(such as foundation depth), and its expression is:

[0074] ;

[0075] in, It is a membership function of a triangle. The vertex of the triangle, , For the left and right expansion points, specifically the numerical shape parameters of the triangle membership function. Taking the foundation depth of a tower as an example, if the design value is 3.0 m, the parameters are set as follows: , , .

[0076] The S-shaped membership function is applicable to gradually changing influence indicators, such as the goaf area ratio, and its expression is:

[0077] ;

[0078] in, It represents the Sigmoid membership degree. The inflection point value is set as the critical goaf area ratio that triggers significant surface subsidence, determined based on the geological report of the mining area. ; The slope parameter is determined by nonlinear least squares fitting of historical data from the mining area.

[0079] (2) Indicator grouping and primary reasoning

[0080] The 36 tertiary indicators were grouped as shown in Table 3, and then input into three independent primary fuzzy inference engines. The results were aggregated into a comprehensive evaluation value for the probability of occurrence (P), severity (S), and detectability (D). Table 3 is the indicator grouping mapping table.

[0081] Table 3

[0082]

[0083] Each primary inference engine employs a hierarchical structure based on secondary indices for aggregation. Specifically, input variables are first grouped according to secondary indices, and fuzzy inference is performed within each group. The output membership vector of each group's inference is defuzzified into a single value using the centroid method, and this single value serves as the input variable for the next level of inference. The centroid method calculation formula is as follows:

[0084] ;

[0085] in, This is the precise value (single value) output after centroid deblurring. For the universe of discourse of the output variable, This is the membership function of the aggregated output.

[0086] Through layer-by-layer aggregation, the final output is a comprehensive score for the target variable (P, S, or D), with a value range of [0, 100]. The number of rules is controlled by grouping the input variables, with each group consisting of 2 to 3 variables, to keep the rule size within the maintainable range of the project, preferably not exceeding 54 rules.

[0087] (3) Advanced fuzzy reasoning

[0088] The P, S, and D scores (all single values ​​within the range [0, 100]) output from the primary reasoning are used as input to the Mamdani advanced reasoning engine. Each of P, S, and D takes three linguistic values ​​(low / medium / high), resulting in a rule base of 3 × 3 × 3 = 27 rules. A typical rule example is shown below:

[0089] IF (P IS high) AND (S IS high) AND (D IS low) THEN (risk IS high).

[0090] The initial risk score is obtained by using a min-max inference mechanism and the centroid method for defuzzification. .

[0091] like Figure 2 As shown, Figure 2 This is a schematic diagram of a hierarchical fuzzy inference architecture. For details of the hierarchical fuzzy inference process described above, please refer to... Figure 2The 36 tertiary indicators are grouped into secondary indicators and then input into the corresponding primary fuzzy inference engines. After hierarchical fuzzy inference and centroid defuzzification, the three comprehensive scores of probability of occurrence (P), severity (S), and detectability (D) are obtained. Then, P, S, and D are input into the advanced Mamdani fuzzy inference engine to obtain the complete reasoning process of the initial risk score. This intuitively reflects the hierarchical reasoning structure of dimensionality reduction of high-dimensional indicators and control of the number of rules.

[0092] Step S4, Disaster Recovery Time Risk Correction Function.

[0093] Step S4, based on the initial risk score, introduces a disaster recovery time factor and calculates a correction coefficient using a risk correction function to obtain a risk score corrected for recovery time. The specific process is as follows:

[0094] Define the recoverable time factor score for geological disasters The Delphi method yields a comprehensive recovery difficulty score, ranging from 0 to 100. The Delphi method involves at least five experts who independently score the assessment, considering factors such as accessibility, geological complexity, resource allocation capabilities, and backup line switching capabilities. The average score is then used as the final score. .

[0095] The correction function is in logarithmic form:

[0096] ;

[0097] in: The comprehensive recovery difficulty score, obtained by the Delphi method, is used to score the recoverable time factor of geological disasters, with a value range of 0 to 100. This is a function to correct the recovery time of geological disasters, and the output is a risk correction coefficient. for The normalization upper limit is fixed at 100; The adjustment coefficient can be calibrated by minimizing the historical risk assessment error, and an initial value of 0.15 is recommended. The higher the value, the more significant the amplification effect of recovery time on risk. Mining areas can use local emergency repair capacity statistics to assess... Optimize and adjust.

[0098] Final risk score The calculation formula is:

[0099] ;

[0100] in, For the initial risk score; This is a risk score adjusted for recovery time.

[0101] like Figure 3 As shown, Figure 3 This is a schematic diagram of the continuous correction function relationship for TTR risk. For details on the correspondence between the correction factors, please refer to [link / reference]. Figure 3 To restore the difficulty rating x-axis, correction factor Plot the graph with the vertical axis as the ordinate. The change curve of the logarithmic continuous correction function when the default value is 0.15 clearly reflects the nonlinear quantitative relationship between post-disaster recovery capacity and risk correction coefficient.

[0102] Example 2, as Figure 4 As shown, Figure 4 This is a flowchart of steps S1-S6 of the comprehensive risk assessment method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making in Embodiment 2 of the present invention. Based on the method in Embodiment 1, the method in Embodiment 2 further includes the following steps:

[0103] Step S5: Weight fusion and conflict self-checking mechanism.

[0104] Step S5 calculates the subjective and objective weights of the indicators, converts the weights into the basic probability distribution of the evidence theory, measures the degree of conflict between the subjective and objective weights and performs fusion or circuit breaking processing, and calibrates the risk score after recovery time correction using the comprehensive weights.

[0105] In step S5, the subjective weights are calculated using the triangular fuzzy number hierarchical analysis method, and the objective weights are calculated using the information entropy method.

[0106] In step S5, the Jussellmey evidence distance metric is used to weigh the conflict degree. When the conflict degree exceeds a preset threshold, circuit breaking is performed and subjective weights are applied.

[0107] In step S5, when converting weights to basic probability allocation, a preset reliability coefficient is used for discounting. The specific process is as follows:

[0108] (1) Calculation of subjective weights

[0109] A triangular fuzzy number based on the Saaty 1-9 scale was used, and pairwise comparisons of the primary indicators were performed by multiple experts. The triangular fuzzy number is represented as follows: ,in As the lower limit, The most likely value, The upper limit is used. The geometric mean method is used to aggregate the group judgment matrix, and the subjective weight vector of each indicator is calculated. .

[0110] in, This is a triangular fuzzy number used to represent the fuzzy judgments made by experts when comparing pairs of objects; This is the lower limit of the triangular fuzzy number; (In the triangular fuzzy number) is the most likely value (peak value) of the triangular fuzzy number; This represents the upper limit of the triangular fuzzy number; This is the subjective weight vector, calculated using the FAHP method; The subjective weight of the nth geological disaster risk indicator; This represents the total number of geological disaster risk indicators used in the calculation.

[0111] (2) Calculation of objective weights

[0112] Calculate the objective weight vector based on the information entropy of data from each potential hazard point. . The objective weight vector is calculated using the entropy weight method. The objective weight of the nth indicator is calculated using the following formula:

[0113] ;

[0114] in, This serves as a traversal index for geological disaster risk indicators. For the first Information entropy of each geological disaster risk indicator.

[0115] Information entropy The calculation formula is:

[0116] ;

[0117] in, The information entropy of the nth geological disaster risk indicator; (In the entropy weight method) represents the total number of potential hazards; For the first The first potential hazard point is at the Normalized values ​​for individual geological disaster risk indicators.

[0118] (3) Weighting-Evidence Transformation

[0119] Subjective weights and objective weights originate from two different information sources: expert perception and data distribution. In engineering, these are considered weakly independent evidence. and These are converted to the Basic Probability Assignment (BPA) in Dempster-Shafer's evidence theory. The conversion formula is:

[0120] ;

[0121] in, For the first Each evaluation level (first-level indicator) corresponds to a specific element. For the complete collection; For reliability, a value of 0.9 is recommended. For Jiao Yuan Basic probability allocation (BPA); The BPA value assigned to the entire set represents the remaining uncertainty; denoted as the weight of the i-th geological disaster risk indicator.

[0122] (4) Conflict measurement and integration

[0123] The Jousselme evidence distance is used to measure the degree of conflict between two BPAs. Let... and The BPA is derived from subjective weights and objective weights, respectively, and the Jousselme distance is also given. The calculation formula is:

[0124] ;

[0125] in, Let be the similarity matrix between focal elements, with elements in, For Jiao Yuan With Jiao Yuan The cardinality (number of elements) of the intersection; For Jiao Yuan With Jiao Yuan The cardinality (number of elements) of the union; For the i-th focal element, that is, the i-th evaluation level; : The j-th focal element, i.e., the j-th evaluation level.

[0126] Normalize the evidence distance to the degree of conflict. :

[0127] ;

[0128] in, This represents the theoretical maximum evidence distance. The value range is [0,1], where 0 represents complete agreement and 1 represents complete conflict. The fusion decision rule is as follows:

[0129] when At that time, the Dempster-Shafer synthesis rule was used for fusion to obtain the comprehensive weight. ;

[0130] when At that time, stop weight fusion and adopt subjective weighting. As an output weight, it also triggers an early warning of conflicts between subjective and objective information, prompting managers to verify sensor data or reorganize expert assessments.

[0131] In this embodiment of the invention, the threshold of 0.4 was determined by sensitivity analysis of historical assessment data from multiple mining areas, and its corresponding confidence level is approximately 95%.

[0132] For the comprehensive weight calibration of risk scores corrected for recovery timeliness, this invention employs a three-layered technical logic: weights permeate the entire hierarchical fuzzy inference process; weights determine the initial risk score; and the initial risk score forms the basis for the corrected score calculation. Specifically, in steps S2 (indicator system construction), S3 (hierarchical fuzzy inference architecture), and S5 (weight fusion and conflict self-checking mechanism), it is clarified that each level of the four-dimensional, three-layered geological disaster risk indicator system must rely on the comprehensive weights obtained through the fusion of subjective and objective factors for fuzzy inference calculation. These comprehensive weights are the initial risk score output by advanced fuzzy inference in step S3, which aggregates the probability of occurrence (P), severity (S), and detectability (D) scores. The core parameters; and the risk correction function for the disaster recovery time in step S4 specifies the corrected risk score for step S4. From the initial risk score The initial risk score is calculated by multiplying by the recoverable time correction factor. It is entirely determined by the comprehensive weight. Therefore, the comprehensive weight controls the core calculation link of hierarchical fuzzy inference and completes the calibration of the risk score in step S4 from the bottom layer.

[0133] Step S6: Dynamically update the triggering mechanism.

[0134] Step S6 monitors dynamic parameters based on preset configurable thresholds. When preset trigger conditions are met, steps S1 to S5 are re-executed based on the calibrated risk score to complete the dynamic update of the entire risk rating process. The specific process is as follows:

[0135] The trigger threshold adopts a configurable strategy, and the default reference values ​​and their basis are shown in Table 4. Table 4 shows the dynamically updated trigger thresholds and their basis.

[0136] Table 4

[0137]

[0138] When any trigger condition is met, the system automatically extracts the latest monitoring data and re-executes the risk rating process for the affected area.

[0139] Using the calibrated risk score as a benchmark, steps S1 to S5 are re-executed to complete the dynamic update of the entire risk rating process. The complete risk rating process is defined as follows: Step S1 – Multi-source data acquisition and dynamic degradation processing; Step S2 – Construction of a four-dimensional, three-layer indicator system; Step S3 – Hierarchical fuzzy reasoning; Step S4 – Correction of disaster recovery time; and Step S5 – Integration of subjective and objective weights and conflict self-checking. This corresponds to “re-executing steps S1 to S5.” Simultaneously, the calibrated risk score output in Step S5 is the final result of the entire risk rating process. The core of the dynamic update mechanism is to dynamically correct the existing risk level of potential hazards in real time. This requires recalculation based on the final result of the previous round of rating and the latest monitoring data. Therefore, “using the calibrated risk score in Step S5 as a benchmark” is an inherent requirement of the core design of this invention for dynamic updating, dynamic correction, and iterative evaluation.

[0140] Example 3 is an experimental example based on the content of Examples 1 and 2.

[0141] Taking the #07 landslide hazard point of a 35kV power supply line L1 in a mining area in North China as an example, the complete calculation process is shown.

[0142] (1) Data collection

[0143] 32 measured indicators were obtained, and 4 indicators were replaced by grade B or C because no measured values ​​were available.

[0144] (2) Indicator fuzzification and primary reasoning

[0145] Taking the slope angle as an example, the measured value The membership degree of each language value is obtained by calculating the trapezoidal membership function: , , , 0.67, That is, the membership vector is .

[0146] After defuzzifying the 18 indicators in group P using fuzzy inference and centroid method within each group, they are aggregated layer by layer, resulting in a final comprehensive score of 65.0 for P. Similarly, the comprehensive score for S is 82.0, and the comprehensive score for D is 48.0.

[0147] (3) Advanced reasoning and TTR correction

[0148] Will Input an advanced fuzzy inference engine and output an initial risk score. 72.5.

[0149] Delphi method to obtain , Using a default value of 0.15, calculate the correction function:

[0150] .

[0151] Final risk score:

[0152] .

[0153] (4) Weight fusion

[0154] Subjective weight vector Objective weight vector The conflict level is calculated using Jousselme distance. Successful integration yields a comprehensive weight. .

[0155] (5) Evaluation conclusions

[0156] The comprehensive risk score for hazard #07 is 77.9, corresponding to Level II (high risk), requiring engineering protective measures to be implemented within one week. In practical engineering applications, some hazard points may experience fluctuations in model output due to missing indicator data or significant conflicts between subjective and objective weights. In such cases, the built-in data degradation mechanism and conflict self-checking circuit breaker mechanism of this invention will automatically intervene to correct the situation, ensuring the robustness of the assessment results under non-ideal data environments.

[0157] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0158] The above embodiments are merely illustrative examples and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A comprehensive risk assessment method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making, characterized in that, The method includes the following steps: Step S1: Collect multi-dimensional indicator data of potential geological hazards in coal mine power supply lines, classify data sources according to confidence level, remove missing indicators and redistribute weights proportionally to obtain a standardized indicator dataset. Step S2: Based on geological background, topography, meteorology and hydrology, and engineering management, construct a first-level, second-level, and third-level geological disaster risk indicator system, and match the standardized indicator dataset to the corresponding third-level indicators; Step S3: The matched three-level indicators are fuzzified using the adaptive membership function. After aggregation by primary fuzzy inference, three scores are obtained: probability of occurrence, severity, and detectability. The three scores are then input into the advanced fuzzy inference engine to output the initial risk score. Step S4: Based on the initial risk score, introduce the geological disaster recovery time factor, calculate the correction coefficient through the risk correction function, and obtain the risk score corrected for recovery time.

2. The comprehensive risk assessment method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making as described in claim 1, characterized in that, The method further includes the following steps: Step S5: Calculate the subjective and objective weights of the indicators, convert the weights into the basic probability distribution of evidence theory, measure the degree of conflict between subjective and objective weights and perform fusion or circuit breaking processing, and calibrate the risk score after recovery time correction using the comprehensive weights. Step S6: Based on the preset configurable threshold, monitor the dynamic parameters. When the preset triggering conditions are met, use the calibrated risk score as a benchmark and re-execute steps S1 to S5 to complete the dynamic update of the risk rating throughout the entire process.

3. The comprehensive risk assessment method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making as described in claim 1, characterized in that, In step S1, the data source is divided into three levels of confidence based on measured data, historical statistical values, and expert experience. The weight of missing indicators is redistributed according to the proportion of remaining indicators belonging to the same higher-level indicator.

4. The comprehensive risk assessment method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making as described in claim 1, characterized in that, The membership functions mentioned in step S3 include three types: trapezoidal, triangular, and S-shaped, which are respectively adapted to the three-level index with standardized threshold, design optimal value, and nonlinear gradual change characteristics.

5. The comprehensive risk assessment method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making as described in claim 1, characterized in that, In step S3, the primary fuzzy inference group and aggregate the three-level indicators into single-value scores, while the advanced fuzzy inference uses the Mandani fuzzy inference mechanism and the centroid method to defuzzify and output the initial risk score.

6. The comprehensive risk assessment method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making as described in claim 1, characterized in that, In step S3, the rule size of the primary fuzzy inference is controlled, while the rule base of the advanced fuzzy inference consists of low / medium / high linguistic values ​​of occurrence probability, severity, and detectability.

7. The comprehensive risk assessment method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making as described in claim 1, characterized in that, The calculation formulas for the risk correction function and the risk score adjusted for recovery time in step S4 are as follows: ； ； In the formula, Scoring the recovery time factor for geological disasters; This is the maximum score. This is the adjustment coefficient; For the initial risk score; This is a risk score adjusted for recovery time.

8. A comprehensive risk assessment method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making, as described in claim 2, is characterized in that... The subjective weights mentioned in step S5 are calculated using the triangular fuzzy number hierarchical analysis method, while the objective weights are calculated using the information entropy method.

9. A comprehensive risk assessment method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making, as described in claim 2, is characterized in that... In step S5, the Jussellmey evidence distance metric is used to weigh the conflict degree. When the conflict degree exceeds a preset threshold, circuit breaking is performed and subjective weights are applied.

10. A comprehensive risk assessment method for geological disaster hazards in coal mine power supply lines based on fuzzy logic and multi-criteria decision-making, as described in claim 2, is characterized in that... In step S5, when converting the weights to basic probability allocation, a preset reliability coefficient is used for discounting.

Images