An intelligent coal mine accident risk assessment method based on model simulation

Abstract

This invention discloses an intelligent coal mine accident risk assessment method based on model simulation, belonging to the field of model simulation technology. The method includes: collecting multi-source risk factor data, hierarchically screening core risk factors, standardizing the data, and outputting a standardized core risk factor set; constructing a risk factor correlation and transmission node library based on the standardized core risk factor set, launching a digital twin simulation model, performing dynamic correlation and transmission deduction, correcting data deviations in real time, and outputting a dynamic scenario state set; constructing a risk simulation evolution step-by-step stage library, matching the initial evolution stage based on the dynamic scenario state set, advancing the step-by-step evolution based on the simulation time scale, recording key data in real time, and forming a dataset of the entire simulation evolution process; and constructing a five-level risk level ladder judgment library, extracting core assessment indicators from the dataset of the entire simulation evolution process, and comprehensively judging the final risk level, significantly improving the accuracy and practicality of coal mine accident risk assessment.

Description

Technical Field

[0001] This invention relates to the field of model simulation technology, specifically to an intelligent coal mine accident risk assessment method based on model simulation. Background Technology

[0002] Coal mine accident risk assessment is a crucial step in ensuring safe coal mine production. Currently, traditional risk assessment methods mainly rely on historical accident statistical analysis, expert experience judgment, and static risk assessment models. These methods are typically based on limited data samples and fixed assessment rules, identifying and classifying potential risks through manual analysis or simple mathematical models. For example, some existing technologies use statistical methods to analyze environmental monitoring data or assign weights to risk factors through expert scoring to calculate a comprehensive risk value. In addition, some studies have attempted to introduce static simulation models to simulate the risk evolution process under specific accident scenarios.

[0003] However, existing technologies have significant limitations. First, traditional methods do not adequately consider the dynamic correlations between multiple risk factors, especially the coupling effects between environmental, equipment, geographical, and human factors. This leads to significant deviations between risk assessment results and actual scenarios. For example, the linkage between abnormal gas concentrations and ventilation equipment failures is often simplified as an independent event, failing to accurately reflect the complexity of accident evolution. Second, existing simulation models are mostly static or semi-dynamic, lacking real-time data correction mechanisms and failing to respond promptly to changes in the underground environment, resulting in a disconnect between simulation results and actual risk conditions. Furthermore, existing risk assessment methods lack systematicity and flexibility in stage division and level determination. The risk evolution process is often simplified to a few stages, and the transition conditions between stages are vaguely defined, failing to accurately match the progressive characteristics of coal mine accidents. At the same time, risk level determination often relies on single indicators or rigid rules, failing to fully consider the dynamic adjustment role of on-site control capabilities, resulting in insufficient practicality and relevance of assessment results. These problems limit the accuracy and reliability of traditional methods in complex coal mine environments. Summary of the Invention

[0004] (a) Technical problems to be solved To address the shortcomings of existing technologies, this invention provides an intelligent coal mine accident risk assessment method based on model simulation. By dynamically correlated and transmitted multiple risk factors, real-time correction of digital twin simulation models, and five-level risk ladder determination, it solves the problems of insufficient consideration of the coupling effect of risk factors, disconnect between simulation results and actual scenarios, ambiguous risk assessment stage division, and rigid level determination in traditional methods, significantly improving the accuracy and practicality of coal mine accident risk assessment.

[0005] (II) Technical Solution To achieve the above objectives, the present invention provides the following technical solution: an intelligent coal mine accident risk assessment method based on model simulation, comprising: Collect multi-source risk factor data, set simulation target adaptation benchmarks, hierarchically screen core risk factors, perform standardization processing, and output a standardized core risk factor set; Based on a standardized set of core risk factors, a risk factor correlation and transmission node library is constructed, a digital twin simulation model is launched, dynamic correlation and transmission deduction is performed, data deviations are corrected in real time, and a dynamic scenario state set is output. A risk simulation evolution step-by-step stage library is constructed. Based on the dynamic scenario state set, the initial evolution stage is matched, the step-by-step evolution is promoted based on the simulation time scale, key data is recorded in real time, and a dataset of the entire simulation evolution process is formed. A five-level risk level tiered judgment library is constructed. Core evaluation indicators are extracted from the dataset of the entire simulation evolution process. The final risk level is determined comprehensively, and an evaluation report is generated and output.

[0006] Furthermore, multi-source risk factor data, including environmental, equipment, geographical, and human-related factor data, as well as correlation factor data of similar coal mine accidents over the past 5-10 years; setting simulation target adaptation benchmarks, including: determining the type of simulation target to be evaluated, which is divided into single-accident simulation targets and multi-accident coupled simulation targets; for single-accident simulation targets, setting single-accident correlation benchmarks and clarifying the core factor categories corresponding to this type of accident; for multi-accident coupled simulation targets, setting multi-accident coupled correlation benchmarks and clarifying the core factor categories of various accidents and the coupling correlation rules between core factors.

[0007] Furthermore, the core risk factors are screened in a tiered manner, including: matching each factor with the set simulation target adaptation benchmark one by one, eliminating factors that are not related to the benchmark, and obtaining a subset of related factors; counting the frequency of occurrence of each factor in the subset of related factors in the corresponding type of historical accidents, eliminating invalid factors with a frequency of 0, and retaining the related effective factors; and dividing the related effective factors into three levels: strong impact factors, medium impact factors, and weak impact factors to form a core risk factor set. Standardization processes are implemented, including: setting three levels of state thresholds for each factor in the core risk factor set: normal state threshold, abnormal warning threshold, and accident critical threshold; and using a unified state representation format of normal / abnormal / critical to convert the collected multi-source risk factor data into corresponding state representations.

[0008] Furthermore, a risk factor correlation and transmission node library is constructed, including: treating each core risk factor as an independent node, binding three core pieces of information within the node: triggering condition, transmission object, and transmission intensity level; the triggering condition is the core risk factor changing from a normal state to an abnormal state; the transmission object is other core risk factors that can be directly affected after the core risk factor becomes abnormal; the transmission intensity level is divided into three levels: strong, medium, and weak, corresponding to changes in the state of the transmission object within 1, 2-3, and 4 or more simulation step units, respectively.

[0009] Furthermore, the dynamic correlation transmission simulation includes: importing a standardized set of core risk factors into a digital twin simulation model, assigning an initial state representation and corresponding data to each node; configuring basic simulation parameters, including the simulation time scale, real-time data synchronization cycle, and simulation termination conditions; monitoring node state changes in real time according to the simulation time scale; when any node reaches its own trigger condition, it becomes the initial transmission source node; retrieving the transmission object and transmission intensity level of the initial transmission source node from the risk factor correlation transmission node library; activating the transmission object nodes in the order of strong level, medium level, and weak level and updating their state representation and corresponding data; performing secondary monitoring on the activated transmission object nodes; if they reach their own trigger condition, they are determined as new transmission source nodes, until no new transmission source nodes are generated or the simulation termination condition is reached.

[0010] Furthermore, during the dynamic correlation and transmission simulation, real-time multi-source risk factor data is received according to the set synchronization period. The real-time multi-source risk factor data is compared with the data of the corresponding nodes in the simulation model. If the absolute value of the data deviation exceeds the preset deviation threshold, the state representation and data of the corresponding nodes in the simulation model are corrected based on the real-time multi-source risk factor data. After the correction is completed, the dynamic correlation and transmission simulation is restarted. After the simulation is completed, the dynamic scenario state set is output, which includes the final state representation, data value and transmission process record of each node.

[0011] Furthermore, the risk simulation evolution step-by-step library is constructed by dividing the entire risk evolution process into 7 progressive stages: safe state stage, initial anomaly stage, mild risk stage, moderate risk stage, severe risk stage, accident critical state stage, and accident occurrence state stage. Each stage is bound to three core parameters: factor state combination, scenario characteristics, and evolution duration threshold. The factor state combination clarifies the state combination requirements for each strong / medium influencing factor, while weak influencing factors are not used as the core basis for stage determination. The scenario characteristics clarify the downhole physical scenario description for the corresponding stage. The evolution duration threshold is defined as the shortest duration that each stage must meet.

[0012] Furthermore, based on the dynamic scenario state set, matching the initial evolution stage includes: extracting the current state representation of each strong / medium impact factor in the dynamic scenario state set, and comparing it one by one with the factor state combination parameters of each stage in the risk simulation evolution step stage library; using the rule of full matching priority and core factor matching as a supplement to determine the initial evolution stage, if the factor state combination is completely consistent, it is directly matched; if only the strong impact factor state combination matches, it is combined with scenario features to assist in the determination. The step-by-step evolution based on the simulation timescale includes: after each step unit of deduction is completed, a persistence condition determination and an advancement condition determination are performed; the persistence condition determination checks whether the current scenario factor state combination meets the requirements of the current evolution stage, and if it does, the current stage is maintained; the advancement condition determination checks whether the current scenario factor state combination meets the requirements of the next stage, and if it does, it automatically advances to the next stage; if neither is met, it returns to the previous evolution stage and its duration is recounted.

[0013] Furthermore, during the step-by-step evolution process, if a risk intervention instruction is received from the coal mine site, an intervention simulation is triggered in the digital twin simulation model: the state of the target factor corresponding to the intervention instruction is corrected according to the preset intervention effect; after the correction is completed, the state combination of each factor is re-checked; if the corrected factor state combination falls back to the duration condition of a previous stage, a stage rollback is triggered, returning to the previous stage and restarting the duration counting of that stage; if several prognostic factor state combinations still meet the conditions of the current stage or the next stage, no rollback is triggered, and the evolution continues; key data are recorded in real time at a frequency of 1 step unit / record, forming a dataset of the entire process of coal mine risk simulation evolution.

[0014] Furthermore, a risk level tiered judgment library is constructed, dividing risk levels into five levels. Each risk level is bound to three rigid matching conditions and one flexible adjustment condition: rigid condition one is the final stage of simulation evolution, rigid condition two is the number of core factors of abnormal states, and rigid condition three is the risk characteristic type. The flexible adjustment condition is the actual on-site control capability of the coal mine. From the full-process dataset of coal mine risk simulation evolution, three core evaluation indicators are extracted: the final stage of simulation evolution, the number of core factors of abnormal states, and the risk characteristic type. The extracted three core evaluation indicators are compared one by one with the rigid matching conditions of the risk level tiered judgment library to match the corresponding preliminary risk level for each indicator. If the three preliminary risk levels are completely consistent, the level is directly used as the candidate risk level. If the three preliminary risk levels are inconsistent, the highest preliminary risk level is used as the benchmark level, and fine-tuning is carried out in combination with the on-site control capability in the flexible adjustment condition. After review, the final risk level is determined.

[0015] (III) Beneficial Effects This invention provides an intelligent coal mine accident risk assessment method based on model simulation, which has the following beneficial effects: (1) By collecting multi-source risk factor data and setting simulation target adaptation benchmarks, the risk factors were accurately screened and standardized, effectively eliminating irrelevant factors and invalid data, improving the relevance and reliability of the data, and screening core risk factors in a hierarchical manner and dividing the impact level, which can focus on key disaster-causing factors and provide high-quality input for subsequent simulations. Standardized processing unified the data representation form, ensuring the standardization and consistency of model input, thereby significantly improving the accuracy and efficiency of risk assessment.

[0016] (2) By constructing a risk factor correlation and transmission node library and launching a digital twin simulation model, the dynamic correlation and transmission of risk factors were realized. It can accurately simulate the coupling effect and chain reaction of multi-source risk factors. The real-time data correction mechanism ensures that the simulation model is synchronized with the actual underground environment and reduces deviation. Through node status monitoring and transmission logic activation, the risk evolution path is intuitively displayed, providing dynamic basis for accident early warning. It significantly improves the real-time performance and accuracy of risk assessment and provides scientific and reliable decision support for coal mine safety management.

[0017] (3) By constructing a risk simulation evolution step-by-step stage library, the risk evolution process is divided into 7 progressive stages, and factor state combinations, scenario characteristics and duration thresholds are bound together, realizing the refined simulation of risk evolution. The initial stage is matched based on the dynamic scenario state set, and the stage duration and advancement conditions are monitored in real time through step-by-step deduction. The stage regression or advancement is dynamically adjusted in combination with intervention measures, which significantly improves the accuracy and controllability of the risk evolution process. Key data are recorded in real time to form a full-process dataset, providing a comprehensive and dynamic basis for risk level determination, and enhancing the reliability and pertinence of risk assessment.

[0018] (4) By constructing a five-level risk level tier judgment library, and combining rigid matching conditions and flexible adjustment conditions, the risk level can be accurately judged. Core evaluation indicators are extracted from the dataset of the entire simulation evolution process. The final stage of the simulation, the number of abnormal factors and the type of risk characteristics are comprehensively considered. Through dynamic fine-tuning of on-site control capabilities, the evaluation results are ensured to be highly consistent with the actual risk level, which improves the scientificity and flexibility of risk judgment. The generated evaluation report contains detailed evidence and targeted intervention suggestions, providing a comprehensive and reliable basis for coal mine safety decision-making and significantly enhancing the pertinence and effectiveness of risk control. Attached Figure Description

[0019] Figure 1 This is a schematic diagram illustrating the steps of the intelligent coal mine accident risk assessment method based on model simulation of the present invention. Figure 2This is a schematic diagram of the intelligent coal mine accident risk assessment method based on model simulation according to the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please see Figures 1-2 This invention provides an intelligent coal mine accident risk assessment method based on model simulation, comprising the following steps: Step 1: Collect multi-source risk factor data, set simulation target adaptation benchmarks, stratify and screen core risk factors, perform standardization processing, and output a standardized core risk factor set. Step one includes the following: Step 101: Collect multi-source risk factor data. Environmental and equipment-related factor data are collected in real time through the coal mine underground IoT monitoring system. Related factor data from similar coal mine accidents over the past 5-10 years are retrieved from the historical accident database. Geographic factor data is obtained from the on-site geological survey report. Human-related factor data is preset based on the simulation scenario requirements of this assessment, forming the original risk factor set. Environmental factors include gas concentration, wind speed, humidity, temperature, and dust concentration; equipment factors include ventilation fan operating power, hydraulic support pressure, conveyor belt operating speed, and water pump drainage; geographic factors include coal seam thickness, roadway dip angle, roof lithology, and hydrogeological type; and human-related factors include operator qualification level, operating procedure compliance rate, and shift frequency. The preset data is determined based on statistical data of accidents caused by human factors in the mine over the past 3 years. Step 102: Set simulation target adaptation benchmarks. Based on the actual mining needs and safety management priorities at the coal mine site, determine the simulation target type for this assessment. Simulation target types are divided into single-accident simulation targets and multi-accident coupled simulation targets. For single-accident simulation targets, set single-accident association benchmarks and clarify the core factor categories corresponding to this type of accident. For example, a gas explosion accident corresponds to environmental and equipment core factors. For multi-accident coupled simulation targets, set multi-accident coupling association benchmarks and clarify the core factor categories of various accidents and the coupling association rules between core factors. For example, for a gas explosion and roof collapse coupled accident, the coupling association rules need to be clarified as follows: the preceding factor is the ventilation fan operating status, the following factor is the roof pressure, the association trigger condition is that the ventilation fan operating power is lower than 80% of the rated power, and the impact logic is: reduced ventilation fan operating power → decreased wind speed → gas accumulation in the roadway → uneven roof stress → increased roof pressure. Step 103: Layered screening of core risk factors. The first step is initial screening, matching each factor in the original risk factor set with the established simulation target adaptation benchmark, eliminating factors unrelated to the benchmark (e.g., when assessing a gas explosion accident, hydrogeological type factors are eliminated), resulting in a subset of related factors. The second step is fine screening, statistically analyzing the frequency of each factor in the related factor subset in corresponding types of historical accidents, eliminating invalid factors with a frequency of 0 in historical accidents, and retaining valid factors directly related to the accident occurrence. The third step is hierarchical division, based on the contribution of historical accidents to disaster prevention. Taking into account the results of on-site safety expert demonstrations, the relevant effective factors are divided into three levels: strong impact factors, medium impact factors, and weak impact factors. Among them, strong impact factors are the key triggering factors for accidents, such as gas concentration and ventilation fan operation status in gas explosion accidents, and are mandatory input factors for digital twin simulation models. Medium impact factors are auxiliary driving factors for accident evolution, such as wind speed and roadway inclination angle. Weak impact factors are indirect influencing factors for the occurrence of accidents, such as the qualification level of workers. Medium and weak impact factors are optional adjustment factors for digital twin simulation models, ultimately forming the core risk factor set. Step 104: Referring to coal mine safety regulations and actual on-site control thresholds, set three-level state thresholds for each factor in the core risk factor set: normal state threshold, abnormal warning threshold, and accident critical threshold; adopt a unified state representation format of normal / abnormal / critical to convert the collected multi-source risk factor data into corresponding state representations. For example, if the original gas concentration data is 0.9%, the corresponding normal state threshold is <1%, and the standardized representation is normal; perform consistency verification on the converted state data, remove abnormal data with incorrect data format and contradictory state representations, and finally output a standardized core risk factor set.

[0022] When using this method, refer to steps 101 to 104: By collecting multi-source risk factor data and setting simulation target adaptation benchmarks, the system achieves accurate screening and standardized processing of risk factors, effectively eliminating irrelevant factors and invalid data, improving the relevance and reliability of the data, and stratifying core risk factors and classifying their impact levels. This allows the system to focus on key disaster-causing factors, providing high-quality input for subsequent simulations. Standardized processing unifies the data representation format, ensuring the standardization and consistency of model inputs, thereby significantly improving the accuracy and efficiency of risk assessment.

[0023] Step 2: Based on the standardized core risk factor set, construct a risk factor correlation and transmission node library, start the digital twin simulation model, perform dynamic correlation and transmission simulation, correct data deviations in real time, and output a dynamic scenario state set; Step two includes the following: Step 201: Based on the standardized core risk factor set, combined with case studies of factor-induced disasters from similar coal mine accidents in the past 5-10 years and the expert opinions of more than 3 coal mine safety assessment personnel, construct a risk factor correlation and transmission node library; each core risk factor is an independent node, and three core pieces of information are bound within the node: first, the triggering condition, i.e., the core risk factor changes from a normal state to an abnormal state, consistent with the set three-level state threshold; second, the transmission object, i.e., other core risk factors that can be directly affected after the core risk factor becomes abnormal, specifying the name of the specific core risk factor; third, the transmission intensity level, divided into three levels: strong, medium, and weak. The strong level corresponds to a change in the state of the transmission object within 1 simulation step unit, the medium level corresponds to a change within 2-3 simulation step units, and the weak level corresponds to a change within 4 or more simulation step units. Step 202: Import the standardized core risk factor set into the digital twin simulation model. Assign an initial state representation and corresponding multi-source risk factor data to each node. Configure the basic simulation parameters: The simulation time scale is set to 1 minute / step unit, which can be finely adjusted within the range of 0.5 to 2 minutes / step unit according to the actual risk evolution speed of the mine. The real-time data synchronization cycle is set to 5 minutes. The synchronization data source is the real-time data collected by the coal mine underground Internet of Things monitoring system. The simulation termination condition is set to no new transmission source node generated and continuous for 3 simulation step units or any core risk factor reaching the accident critical threshold. Step 203: Start the digital twin simulation model to perform dynamic correlation and transmission simulation of core risk factors. Monitor the status changes of each node in real time according to the set simulation time scale: when the status of any node reaches its own trigger condition, it turns into an abnormal state and is determined to be the initial transmission source node; retrieve the transmission object and transmission intensity level of the initial transmission source node from the risk factor correlation and transmission node library, and activate the transmission object nodes in the order of strong level, medium level, and weak level. Modify the status representation of the transmission object node and the corresponding multi-source risk factor data according to the time rules corresponding to the transmission intensity level; perform secondary monitoring on the activated transmission object node. If its status reaches its own trigger condition, it is determined to be a new transmission source node. Repeat the above transmission logic until no new transmission source node is generated or the simulation termination condition is reached. Step 204: During the correlation and transmission simulation, receive real-time multi-source risk factor data according to the set 5-minute synchronization cycle; compare the real-time multi-source risk factor data with the data of the corresponding nodes in the digital twin simulation model. If the absolute value of the data deviation exceeds the preset deviation threshold (default is 5%), and the deviation is calculated as |real-time data - simulation data| / real-time data × 100%, then the state representation of the corresponding nodes and the multi-source risk factor data in the digital twin simulation model are corrected based on the real-time multi-source risk factor data; after the correction is completed, restart the dynamic correlation and transmission simulation in step 203 to ensure the consistency between the simulation scenario and the actual underground coal mine scenario; after the simulation is completed, output the dynamic scenario state set after factor coupling and transmission, including the final state representation, data value, and transmission process record of each node.

[0024] When using this method, refer to steps 201 to 204: By constructing a risk factor correlation and transmission node library and launching a digital twin simulation model, the dynamic correlation and transmission of risk factors were realized. This model can accurately simulate the coupling effect and chain reaction of multi-source risk factors. The real-time data correction mechanism ensures that the simulation model is synchronized with the actual underground environment, reducing deviations. Through node status monitoring and transmission logic activation, the risk evolution path is intuitively displayed, providing dynamic basis for accident early warning. This significantly improves the real-time performance and accuracy of risk assessment and provides scientific and reliable decision support for coal mine safety management.

[0025] Step 3: Construct a risk simulation evolution step-by-step stage library. Based on the dynamic scenario state set, match the initial evolution stage, promote the step-by-step evolution based on the simulation time scale, record key data in real time, and form a dataset of the entire simulation evolution process. Step three includes the following: Step 301: Referring to the progressive nature of coal mine accidents and the relevant requirements of the guidelines for the construction of a dual prevention mechanism for coal mine safety risk classification and control and hidden danger investigation and management, and combining the results of expert demonstrations by more than 3 people with coal mine safety assessment experience, a risk simulation evolution step-by-step stage library is constructed. The entire risk evolution process is divided into 7 progressive stages, namely: safe state stage, initial anomaly stage, mild risk stage, moderate risk stage, severe risk stage, accident critical state stage, and accident occurrence state stage. Three core parameters are bound to each stage: First, factor state combination, which clarifies the normal / abnormal / critical state combination requirements of each strong / medium influencing factor, and weak influencing factors are not used as the core basis for stage determination; Second, scene characteristics, which clarifies the underground physical scene description of the corresponding stage, such as local gas accumulation, cracks in the roadway roof, etc.; Third, evolution duration threshold, which is the minimum duration that each stage must meet, in the unit of simulation step unit, such as ≥5 step units for the safe state stage and ≥3 step units for the initial anomaly stage. Step 302: Based on the dynamic scenario state set, extract the current state representation of each strong / medium impact factor and compare it one by one with the factor state combination parameters of each stage in the risk simulation evolution step stage library; use the rule of full matching priority and core factor matching as a supplement to determine the initial evolution stage: if the factor state combination of the dynamic scenario state set is completely consistent with the factor state combination of any stage, then directly match that stage; if they are not completely consistent, only the strong impact factor state combination is matched, then combine the scenario features to assist in the determination, finally determine the initial evolution stage corresponding to the current simulation scenario, and record the matching basis; Step 303: Based on the set 1 minute / step unit, and consistent with the simulation timescale, initiate step-by-step evolution. After each step unit is completed, perform two core judgments: First, the continuity condition judgment checks whether the current scenario's factor state combination still meets the factor state combination requirements of the current evolution stage. If it does, maintain the current stage and synchronously update the data of each factor and scenario feature details, such as the specific numerical changes in gas concentration in the low-risk stage. Second, the advancement condition judgment checks whether the current scenario's factor state combination meets the factor state combination requirements of the next evolution stage. If it does, automatically advance to the next stage and synchronously update the scenario features and the evolution duration count of this stage. If neither the continuity condition of the current stage nor the advancement condition of the next stage is met, return to the previous evolution stage and recount the duration of this stage. Step 304: During the step-by-step evolution process, if a risk intervention instruction is received from the coal mine site, such as turning on the backup ventilation fan, activating the roof reinforcement device, or evacuating workers, the instruction must specify the target factor for the intervention. In this case, the intervention simulation is triggered in the simulation model: the state of the target factor corresponding to the intervention measure is corrected according to the preset intervention effect. For example, after turning on the backup ventilation fan, the wind speed factor state is corrected from abnormal to normal. After the correction is completed, the state combination of each factor is rechecked. If the corrected factor state combination falls back to the duration condition of a previous stage, a stage rollback is triggered, returning to the previous stage and restarting the duration count of that stage. If several prognostic factor state combinations still meet the conditions of the current stage or the next stage, a rollback is not triggered, and the evolution continues. Step 305: Throughout the entire process of the step-by-step evolution simulation, key data is recorded in real time at a frequency of 1 step unit / record. The recorded content includes: the name of the current evolution stage and the duration of the stage, the real-time status representation and corresponding data values ​​of each core risk factor, the description of the current scenario characteristics, whether intervention measures are triggered, and the basis for determining stage advancement / regression. If triggered, the name of the intervention measure, the intervention time, the target factor, and the intervention effect must be recorded. All records are organized in chronological order to form a complete dataset of the entire coal mine risk simulation evolution process.

[0026] When using this method, refer to steps 301 to 305: By constructing a risk simulation evolution step-by-step stage library, the risk evolution process is divided into 7 progressive stages, and factor state combinations, scenario characteristics, and duration thresholds are bound together. This enables refined simulation of risk evolution. The initial stage is matched based on a dynamic scenario state set, and the stage duration and advancement conditions are monitored in real time through step-by-step deduction. Combined with intervention measures, the stage regression or advancement is dynamically adjusted, which significantly improves the accuracy and controllability of the risk evolution process. Key data is recorded in real time to form a full-process dataset, providing a comprehensive and dynamic basis for risk level determination and enhancing the reliability and pertinence of risk assessment.

[0027] Step 4: Construct a five-level risk level tiered judgment library, extract core evaluation indicators from the simulation evolution process dataset, comprehensively determine the final risk level, generate an evaluation report and output it.

[0028] Step four includes the following: Step 401: Based on the relevant clauses of risk classification in the Coal Mine Safety Regulations and the Basic Requirements and Scoring Methods for Coal Mine Safety Production Standardization, and combined with the demonstration results of more than 3 experts with experience in coal mine safety assessment, construct a risk level ladder judgment library; divide the risk level into five levels, in the following order: Level 1 (no risk), Level 2 (low risk), Level 3 (medium risk), Level 4 (high risk), and Level 5 (extremely high risk); bind three rigid matching conditions and one flexible adjustment condition to each level: Rigid condition 1 is the final stage of simulation evolution, such as Level 5 corresponding to the accident occurrence stage, and Level 4 corresponding to the severe risk / accident critical state stage; Rigid condition 2 is the number of core factors of abnormal state, only counting strong / medium impact factors, such as Level 4 requiring ≥3 strong impact factor anomalies or ≥4 medium / strong impact factor combination anomalies; Rigid condition 3 is the risk characteristic type, single factor anomaly / multi-factor coupling anomaly, such as Level 5 requiring multi-factor coupling anomalies; The flexible adjustment condition is the actual control capability of the coal mine on site, including three sub-items: existing intervention measures reserves, emergency response capability of operators, and equipment operation and maintenance support level; Step 402: Extract three core evaluation indicators from the coal mine risk simulation evolution dataset according to the following rules: First, the final stage of the simulation evolution, extracting the name of the final stage at the end of the simulation. If the simulation termination condition is not met, the stage that lasts the longest is taken as the final stage. Second, the number of core factors of abnormal states, counting the total number of strong / medium influencing factors that are in abnormal or critical states at the end of the simulation. Weak influencing factors are not included in the statistics. Third, the risk characteristic type, comparing the factor state combination of the final stage. If only a single strong / medium influencing factor is abnormal, it is judged as a single factor abnormality. If there are two or more strong / medium influencing factors abnormal, it is judged as a multi-factor coupling abnormality. After extraction, an indicator list is formed, indicating the extraction source of each indicator, such as a certain data coming from the record of the Xth step unit. Step 403: Comprehensively determine the risk level. The first step is preliminary judgment, comparing the three extracted core assessment indicators with the rigid matching conditions of the risk level ladder judgment library one by one to match the corresponding preliminary risk level for each indicator. For example, if the final stage is a severe risk stage, the preliminary level is matched as level four; if there are 3 abnormal strong / medium impact factors, the preliminary level is matched as level four; if there are multiple factor coupling anomalies, the preliminary level is matched as level four. The second step is re-judgment. If the three preliminary risk levels are completely consistent, the level is directly used as the candidate risk level; if the three preliminary levels are inconsistent, the highest preliminary risk level is used as the benchmark level, and fine-tuning is made in combination with the on-site control capabilities in the flexible adjustment conditions: if the on-site control capabilities meet the intervention requirements corresponding to the benchmark level, such as level four requiring 3 or more targeted intervention measures, the benchmark level is downgraded by one level as the candidate level; if not, the benchmark level remains unchanged. The third step is final judgment, organizing 2 or more coal mine safety experts to review the candidate level. If the review is passed, it is determined as the final risk level; if it is not passed, the indicator extraction and preliminary / re-judgment process are re-verified. Step 404: Generate an assessment report. The report must include: the final risk level and the basis for judgment, listing the key data for the initial and reassessments; the risk type, specifying whether it is a single-accident risk or a multi-accident coupled risk, and for multi-accident coupled risks, the sub-risk levels of each type of single accident and their coupling effects must be marked; the scope of risk impact, determined based on the characteristics of the simulation scenario, such as a 100-meter range in the middle section of roadway #3; targeted intervention recommendations, written in the format of evolution stage, intervention measures, target factors, and expected effects, such as turning on the backup ventilation fan in the mild risk stage, with wind speed as the target factor and expected effects of wind speed returning to normal and gas concentration decreasing; a deviation analysis between the simulation and the actual scenario, comparing the simulated data with real-time underground data, explaining the reasons for the deviation and the correction measures; simultaneously outputting the assessment report to the coal mine safety management platform and the handheld terminals of underground workers; integrating the final risk level, assessment report, evolution process dataset, and judgment process records to form an assessment archive, which is then stored in the coal mine safety assessment historical database.

[0029] When using this method, please refer to the content of steps 401 to 404: By constructing a five-level risk level tiered judgment library and combining rigid matching conditions with flexible adjustment conditions, accurate risk level determination was achieved. Core evaluation indicators were extracted from the simulation evolution process dataset, comprehensively considering the final stage of the simulation, the number of abnormal factors, and the type of risk characteristics. Dynamic fine-tuning through on-site control capabilities ensured that the evaluation results highly matched the actual risk level, improving the scientific nature and flexibility of risk judgment. The generated evaluation report contains detailed evidence and targeted intervention suggestions, providing a comprehensive and reliable basis for coal mine safety decision-making and significantly enhancing the pertinence and effectiveness of risk control.

[0030] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, and combinations thereof. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0031] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0032] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. An intelligent coal mine accident risk assessment method based on model simulation, characterized in that: include: Collect multi-source risk factor data, set simulation target adaptation benchmarks, hierarchically screen core risk factors, perform standardization processing, and output a standardized core risk factor set; Based on a standardized set of core risk factors, a risk factor correlation and transmission node library is constructed, a digital twin simulation model is launched, dynamic correlation and transmission deduction is performed, data deviations are corrected in real time, and a dynamic scenario state set is output. A risk simulation evolution step-by-step stage library is constructed. Based on the dynamic scenario state set, the initial evolution stage is matched, the step-by-step evolution is promoted based on the simulation time scale, key data is recorded in real time, and a dataset of the entire simulation evolution process is formed. A five-level risk level tiered judgment library is constructed. Core evaluation indicators are extracted from the dataset of the entire simulation evolution process. The final risk level is determined comprehensively, and an evaluation report is generated and output.

2. The intelligent coal mine accident risk assessment method based on model simulation according to claim 1, characterized in that: Multi-source risk factor data, including environmental, equipment, geographical, and human factors, as well as data on factors related to similar coal mine accidents over the past 5 to 10 years; Setting simulation target adaptation benchmarks includes: determining the type of simulation target to be evaluated, which is divided into single-accident simulation targets and multi-accident coupled simulation targets; for single-accident simulation targets, setting single-accident correlation benchmarks and clarifying the core factor categories corresponding to this type of accident; for multi-accident coupled simulation targets, setting multi-accident coupling correlation benchmarks and clarifying the core factor categories of various types of accidents and the coupling correlation rules between core factors.

3. The intelligent coal mine accident risk assessment method based on model simulation according to claim 2, characterized in that: The core risk factors are screened in a tiered manner, including: matching each factor with the set simulation target adaptation benchmark one by one, eliminating factors that are not related to the benchmark, and obtaining a subset of related factors; counting the frequency of occurrence of each factor in the subset of related factors in the corresponding type of historical accidents, eliminating invalid factors with a frequency of 0, and retaining the related effective factors; and dividing the related effective factors into three levels: strong impact factors, medium impact factors, and weak impact factors to form a core risk factor set. Standardization processes are implemented, including: setting three levels of state thresholds for each factor in the core risk factor set: normal state threshold, abnormal warning threshold, and accident critical threshold; and using a unified state representation format of normal / abnormal / critical to convert the collected multi-source risk factor data into corresponding state representations.

4. The intelligent coal mine accident risk assessment method based on model simulation according to claim 1, characterized in that: A risk factor correlation and transmission node library is constructed, including: treating each core risk factor as an independent node, binding three core pieces of information within the node: trigger condition, transmission object, and transmission intensity level; the trigger condition is the core risk factor changing from a normal state to an abnormal state; the transmission object is other core risk factors that can be directly affected after the core risk factor becomes abnormal; the transmission intensity level is divided into three levels: strong, medium, and weak, corresponding to changes in the state of the transmission object within 1, 2-3, and 4 or more simulation step units, respectively.

5. The intelligent coal mine accident risk assessment method based on model simulation according to claim 4, characterized in that: The dynamic correlation transmission simulation includes: importing a standardized set of core risk factors into a digital twin simulation model, assigning an initial state representation and corresponding data to each node; configuring basic simulation parameters, including the simulation time scale, real-time data synchronization cycle, and simulation termination conditions; monitoring node state changes in real time according to the simulation time scale; when any node reaches its own trigger condition, it becomes the initial transmission source node; retrieving the transmission object and transmission intensity level of the initial transmission source node from the risk factor correlation transmission node library; activating the transmission object nodes in the order of strong level, medium level, and weak level and updating their state representation and corresponding data; performing secondary monitoring on the activated transmission object nodes; if they reach their own trigger condition, they are determined as new transmission source nodes, until no new transmission source nodes are generated or the simulation termination condition is reached.

6. The intelligent coal mine accident risk assessment method based on model simulation according to claim 5, characterized in that: During the dynamic correlation and transmission simulation, real-time multi-source risk factor data is received according to the set synchronization period. The real-time multi-source risk factor data is compared with the data of the corresponding nodes in the simulation model. If the absolute value of the data deviation exceeds the preset deviation threshold, the real-time multi-source risk factor data shall be used as the standard, and the state representation and data of the corresponding nodes in the simulation model shall be corrected. After the correction is completed, the dynamic correlation and transmission simulation is restarted. After the simulation is completed, the dynamic scenario state set is output, which includes the final state representation, data value and transmission process record of each node.

7. The intelligent coal mine accident risk assessment method based on model simulation according to claim 1, characterized in that: The risk simulation evolution step-by-step stage library includes: dividing the entire risk evolution process into 7 progressive stages, namely, the safe state stage, the initial anomaly stage, the mild risk stage, the moderate risk stage, the severe risk stage, the accident critical state stage, and the accident occurrence state stage; binding three core parameters to each stage: factor state combination, scenario characteristics, and evolution duration threshold; the factor state combination clarifies the state combination requirements for each strong / medium influencing factor, while weak influencing factors are not used as the core basis for stage determination; the scenario characteristics clarify the downhole physical scenario description for the corresponding stage; and the evolution duration threshold is defined as the shortest duration that each stage must meet.

8. The intelligent coal mine accident risk assessment method based on model simulation according to claim 7, characterized in that: Based on the dynamic scenario state set, the matching of the initial evolution stage includes: extracting the current state representation of each strong / medium impact factor in the dynamic scenario state set, and comparing it one by one with the factor state combination parameters of each stage in the risk simulation evolution step stage library; the initial evolution stage is determined by the rule of full matching first and core factor matching as a supplement. If the factor state combination is completely consistent, it is directly matched. If only the strong impact factor state combination matches, the scene features are combined to assist in the determination. The step-by-step evolution based on the simulation timescale includes: after each step unit of deduction is completed, a persistence condition determination and an advancement condition determination are performed; the persistence condition determination checks whether the current scenario factor state combination meets the requirements of the current evolution stage, and if it does, the current stage is maintained; the advancement condition determination checks whether the current scenario factor state combination meets the requirements of the next stage, and if it does, it automatically advances to the next stage; if neither is met, it returns to the previous evolution stage and its duration is recounted.

9. The intelligent coal mine accident risk assessment method based on model simulation according to claim 8, characterized in that: During the step-by-step evolution process, if a risk intervention instruction is received from the coal mine site, an intervention simulation is triggered in the digital twin simulation model: the state of the target factor corresponding to the intervention instruction is corrected according to the preset intervention effect; after the correction is completed, the state combination of each factor is re-checked; if the corrected factor state combination falls back to the duration condition of a previous stage, a stage rollback is triggered, returning to the previous stage and restarting the duration counting of that stage; if several prognostic factor state combinations still meet the conditions of the current stage or the next stage, no rollback is triggered, and the evolution continues; key data are recorded in real time at a frequency of 1 step unit / record, forming a dataset of the entire process of coal mine risk simulation evolution.

10. The intelligent coal mine accident risk assessment method based on model simulation according to claim 1, characterized in that: A risk level tiered judgment library is constructed, dividing risk levels into five levels. Each risk level is bound to three rigid matching conditions and one flexible adjustment condition: the first rigid condition is the final stage of simulation evolution, the second rigid condition is the number of core factors of abnormal states, and the third rigid condition is the type of risk characteristic. The flexible adjustment condition is the actual control capability of the coal mine on-site. From the full-process dataset of coal mine risk simulation evolution, three core evaluation indicators are extracted: the final stage of simulation evolution, the number of core factors of abnormal states, and the type of risk characteristic. The three extracted core evaluation indicators are compared one by one with the rigid matching conditions of the risk level tiered judgment library to match the corresponding preliminary risk level for each indicator. If the three preliminary risk levels are completely consistent, the level is directly used as the candidate risk level. If the three preliminary risk levels are inconsistent, the highest preliminary risk level shall be used as the base level, and fine-tuning shall be carried out in combination with the on-site control capabilities in the flexible adjustment conditions. The final risk level shall be determined after review.

Images