Gis-based coal mine safety production intelligent integrated dispatching management system and method

Abstract

The present application relates to the technical field of coal mine safety production, and specifically discloses a GIS-based intelligent integrated dispatching management system and method for coal mine safety production, which collects multi-source safety production data underground and performs spatial correlation; generates a device health degree vector through feature extraction of time sequence signals; couples the health degree vector with personnel positioning information and spatial topology, and calculates risk space probability distribution by using a probability graph; generates a gradientized optimal dispatching strategy by maximum expected utility calculation according to the distribution and a predefined utility function; dynamically generates a risk heat map based on the strategy and GIS spatial analysis, identifies key areas, generates and simulates verification of device control instructions and personnel hierarchical evacuation channels, and finally executes and forms a closed-loop feedback; the present application realizes intelligent integration of risk accurate judgment, strategy dynamic optimization and spatial closed-loop dispatching, and improves the synergy and response accuracy of coal mine safety dispatching.

Description

Technical Field

[0001] This invention relates to the field of coal mine safety production technology, specifically to a GIS-based intelligent integrated scheduling and management system and method for coal mine safety production. Background Technology

[0002] With the advancement of intelligent construction in coal mines, technologies such as Geographic Information Systems (GIS), the Internet of Things (IoT), and big data have been initially applied in mine safety production management. Existing technologies typically integrate various monitoring data (such as gas concentration, equipment status, and personnel location) based on GIS platforms, achieving preliminary "one-map" visual monitoring.

[0003] However, existing technologies have the following shortcomings:

[0004] In existing coal mine safety monitoring and dispatching systems, the risk perception, analysis and decision-making and dispatching control links are isolated from each other, which makes it impossible to automatically generate and execute precise and coordinated spatial dispatching instructions based on the dynamic and complex spatiotemporal risk situation underground. Summary of the Invention

[0005] The purpose of this invention is to provide a GIS-based intelligent integrated scheduling and management system and method for coal mine safety production, in order to solve the problems mentioned above.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] The GIS-based intelligent integrated scheduling and management method for coal mine safety production includes the following steps:

[0008] S1: Collect multi-source safety production data from underground coal mines. Multi-source safety production data includes environmental monitoring data, equipment operating parameters, and personnel location information linked by spatial location.

[0009] S2: The system integrates and extracts features from multi-source safety production data. Based on equipment operating parameters and associated environmental monitoring data, it generates a health vector representing the overall operating status of the equipment by mapping time-series signals to a high-dimensional feature space.

[0010] S3: Couple the health vector with the corresponding personnel location information and spatial topology relationship for analysis. Through the inference process based on the probabilistic graph structure, calculate the probability distribution of different levels of safety risks caused by the equipment under the conditions of spatial location and personnel distribution.

[0011] S4: Based on the probability distribution and the utility functions of different predefined scheduling strategies, the optimal gradient scheduling strategy is generated for the equipment by maximizing the expected utility. The gradient scheduling strategy includes multiple ordered response levels from routine maintenance to emergency shutdown.

[0012] S5: Based on the gradient scheduling strategy and GIS spatial analysis function, it generates and executes specific spatial scheduling instructions, including control instructions for related equipment, evacuation route planning for affected personnel, and adjustment plans for production tasks, forming a closed-loop management from risk perception to control execution.

[0013] As a further aspect of the present invention: S2 specifically includes:

[0014] Adaptive decomposition of time-series signals of equipment operating parameters and associated environmental monitoring data yields multiple intrinsic mode components with different time scales;

[0015] For each intrinsic mode component, calculate the intrinsic mode component energy entropy, sample entropy and zero crossing rate to form the feature set of the intrinsic mode component;

[0016] The feature sets corresponding to multiple intrinsic mode components are weighted and concatenated according to the component time scale to obtain a high-dimensional intermediate feature vector.

[0017] The intermediate feature vector is normalized by its maximum and minimum values, and the processed vector is used as the health vector representing the overall operating status of the equipment.

[0018] As a further aspect of the present invention: the process for obtaining the intrinsic modal components is as follows:

[0019] Preset the number of initial modes and center frequency parameters required for variational mode decomposition;

[0020] By introducing envelope complexity as an optimization objective, the initial number of modes and center frequency parameters are iteratively optimized to adaptively determine the number of modes required for decomposition and the center frequency of each mode.

[0021] Based on the number of modes required for the optimized decomposition and the center frequency of each mode, variational mode decomposition is performed on the time-series signal to obtain multiple eigenmode components with different time scales.

[0022] As a further aspect of the present invention: S3 specifically includes:

[0023] Based on spatial topology, a multi-level risk propagation graph is constructed, including device nodes, spatial region nodes, and global risk nodes. Device nodes and spatial region nodes are connected through spatial proximity, spatial region nodes are connected through topological connectivity, and spatial region nodes and global risk nodes are connected through influence weights.

[0024] The health vector is converted into the initial risk value of the device node, and the personnel density influence factor of each spatial area node is calculated based on the personnel location information.

[0025] The initial risk value is integrated with the personnel density influencing factor and used as the input for each spatial region node. The probability propagation is carried out through multiple rounds of iterations through the risk propagation map until the risk probability value of each node converges.

[0026] Output the risk probability values ​​of global risk nodes and nodes in each spatial region as the probability distribution of different levels of security risks.

[0027] As a further aspect of the present invention: the calculation process of the personnel density influence factor is as follows:

[0028] The physical boundaries of each spatial region node are determined based on the spatial topology, and the number of people located within the physical boundaries in real time is counted to calculate the basic static personnel density of the spatial region.

[0029] Based on the historical sequence of personnel location information within a preset time window, the frequency and direction of personnel crossing the boundaries of adjacent spatial area nodes are calculated to quantify the intensity of personnel dynamic mobility in the spatial area.

[0030] The basic static population density and the intensity of dynamic population mobility are weighted and fused together, and then multiplied by a correction coefficient related to the criticality level of spatial area nodes in the escape route to obtain the population density influence factor.

[0031] As a further aspect of the present invention: S4 specifically includes:

[0032] Construct a two-dimensional evaluation matrix. The row dimension of the two-dimensional evaluation matrix corresponds to different risk levels in the probability distribution, and the column dimension corresponds to the scheduling strategies of multiple predefined ordered response levels.

[0033] For each cell in the evaluation matrix, the product of the probability of occurrence of the risk level and the utility value of the corresponding scheduling strategy under the risk level is calculated to obtain the preliminary expected utility value, and the preliminary expected utility values ​​of all cells are normalized.

[0034] The normalized expected utility value of the cells corresponding to high-risk levels in the evaluation matrix is ​​subject to decay correction;

[0035] Calculate the sum of the corrected expected utility values ​​for each column of scheduling policies, and determine the scheduling policy corresponding to the column with the highest sum as the optimal gradient scheduling policy at the current time.

[0036] As a further aspect of the present invention: the attenuation correction of the normalized expected utility value of the cell corresponding to the high-risk level in the evaluation matrix specifically includes:

[0037] Based on the current production stage and shift information, determine the dynamic risk sensitivity level and obtain the corresponding baseline attenuation coefficient accordingly.

[0038] Based on the level of risk, different level adjustment weights are assigned to the rows of different risk levels in the assessment matrix;

[0039] Based on the historical accident frequency data of the corresponding spatial area within the most recent preset number of days, a historical correction factor is calculated. The historical correction factor is positively correlated with the accident frequency of the same historical period.

[0040] The product of the baseline decay factor, the grade correction weight, and the historical correction factor is used as the final dynamic decay factor to multiply and correct the normalized expected utility value of the corresponding cell.

[0041] As a further aspect of the present invention: S5 specifically includes:

[0042] Based on probability distribution, a risk heat map is dynamically generated in GIS, centered on the equipment, reflecting the spatial spread trend of different safety risk levels;

[0043] Based on the distribution range and intensity of the risk heat map, combined with spatial topological relationships, the equipment sets requiring immediate intervention and the affected associated roadway areas are identified and delineated.

[0044] For the equipment set, based on the level of the gradient scheduling strategy, a differentiated sequence of equipment collaborative control instructions is generated. At the same time, based on the risk gradient of the risk heat map and personnel location information, a graded evacuation channel is planned for personnel in the associated roadway area.

[0045] In GIS, simulate the sequence of coordinated control commands for equipment and hierarchical evacuation routes. After verifying spatial feasibility, issue the commands for execution and update the risk heat map based on real-time data after execution to complete closed-loop verification.

[0046] As a further aspect of the present invention: the process of generating the risk heatmap is as follows:

[0047] Using the spatial location of the equipment as the thermal center point, the probability values ​​of different safety risk levels in the probability distribution are mapped to basic thermal values ​​with different initial intensities.

[0048] Based on the orientation of the tunnel connections and ventilation network in the spatial topology, the spatial propagation of the basic thermal value along the tunnel network is calculated.

[0049] By incorporating the influence factor of population density, the thermal values ​​propagated to various spatial locations are dynamically enhanced;

[0050] The heat values, after propagation and enhancement, are rendered in GIS as a risk heat map with a continuous color gradient.

[0051] The GIS-based intelligent integrated dispatch and management system for coal mine safety production includes:

[0052] The multi-source safety production data acquisition module is used to collect multi-source safety production data in coal mines. The multi-source safety production data includes environmental monitoring data, equipment operating parameters and personnel positioning information that are spatially correlated.

[0053] The equipment health status feature extraction and vectorization module integrates and extracts features from multi-source safety production data. Based on equipment operating parameters and associated environmental monitoring data, it generates a health vector representing the overall operating status of the equipment by mapping time-series signals to a high-dimensional feature space.

[0054] The risk probability inference module coupled with spatial topology performs coupled analysis with the health vector and the corresponding personnel location information and spatial topology relationship. Through the inference process based on the probability graph structure, it calculates the probability distribution of different levels of safety risks caused by the equipment under the conditions of spatial location and personnel distribution.

[0055] The gradient scheduling strategy decision optimization module generates the optimal gradient scheduling strategy for the equipment by calculating the expected utility maximization based on the probability distribution and the utility functions of different predefined scheduling strategies. The gradient scheduling strategy includes multiple ordered response levels from routine maintenance to emergency shutdown.

[0056] The GIS spatial scheduling instruction generation and closed-loop execution module, based on the gradient scheduling strategy and GIS spatial analysis function, generates and executes specific spatial scheduling instructions. The spatial scheduling instructions include control instructions for related equipment, evacuation route planning for affected personnel, and adjustment plans for production tasks, forming a closed-loop management from risk perception to control execution.

[0057] The beneficial effects of this invention are:

[0058] (1) This invention achieves a leap from qualitative judgment to quantitative assessment of safety risks by coupling and analyzing the equipment health vector, real-time personnel distribution, and underground spatial topology, and performing dynamic risk inference based on a probabilistic graphical structure. This method can accurately identify the spatial and hierarchical risk probabilities that equipment anomalies may cause under different personnel distribution and roadway connectivity conditions, thereby overcoming the limitations of traditional monitoring systems that rely solely on single threshold alarms, improving the accuracy and foresight of risk warnings, providing a scientific basis for taking differentiated and precise scheduling interventions, and effectively avoiding false alarms, missed alarms, or over-response caused by incomplete information.

[0059] (2) Based on a gradient scheduling strategy and GIS spatial analysis, this invention achieves a closed-loop linkage of "strategy-instruction-space". The system can automatically generate and simulate verification of equipment collaborative control instructions and personnel graded evacuation paths according to the risk level, ensuring the spatial feasibility and coordination of scheduling instructions. This changes the traditional situation where each subsystem (such as monitoring, control, and evacuation) operates independently and instructions may conflict, forming an integrated closed-loop management of intelligent risk perception, dynamic strategy optimization, precise spatial execution, and real-time effect feedback, improving the overall integrity, coordination, and response efficiency of coal mine safety production scheduling. Attached Figure Description

[0060] The invention will now be further described with reference to the accompanying drawings.

[0061] Figure 1 This is a flowchart of the method of the present invention;

[0062] Figure 2 This is a system block diagram of the present invention. Detailed Implementation

[0063] 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.

[0064] Please see Figure 1 As shown, this invention is a GIS-based intelligent integrated scheduling and management method for coal mine safety production, comprising the following steps:

[0065] S1: Collect multi-source safety production data from underground coal mines. Multi-source safety production data includes environmental monitoring data, equipment operating parameters, and personnel location information linked by spatial location.

[0066] S2: The system integrates and extracts features from multi-source safety production data. Based on equipment operating parameters and associated environmental monitoring data, it generates a health vector representing the overall operating status of the equipment by mapping time-series signals to a high-dimensional feature space.

[0067] S3: Couple the health vector with the corresponding personnel location information and spatial topology relationship for analysis. Through the inference process based on the probabilistic graph structure, calculate the probability distribution of different levels of safety risks caused by the equipment under the conditions of spatial location and personnel distribution.

[0068] S4: Based on the probability distribution and the utility functions of different predefined scheduling strategies, the optimal gradient scheduling strategy is generated for the equipment by maximizing the expected utility. The gradient scheduling strategy includes multiple ordered response levels from routine maintenance to emergency shutdown.

[0069] S5: Based on the gradient scheduling strategy and GIS spatial analysis function, it generates and executes specific spatial scheduling instructions, including control instructions for related equipment, evacuation route planning for affected personnel, and adjustment plans for production tasks, forming a closed-loop management from risk perception to control execution.

[0070] In S1, multi-source safety production data is collected from underground coal mines. This multi-source safety production data includes environmental monitoring data linked by spatial location, equipment operating parameters, and personnel location information, specifically including:

[0071] This step involves the collection of multi-source safety production data in underground coal mines. Specifically, environmental monitoring data is collected by deploying gas, carbon monoxide, temperature, wind speed, and dust sensors in key underground areas such as mining faces, return airways, and electromechanical chambers. These sensors transmit monitoring values ​​to the ground data center at a fixed frequency via an industrial Ethernet ring network or fieldbus.

[0072] The acquisition of equipment operating parameters is accomplished through programmable logic controllers or dedicated monitoring sensors connected to key equipment such as coal mining machines, ventilation fans, drainage pumps, and conveyor belts. The acquired parameters include, but are not limited to, equipment voltage, current, operating speed, vibration amplitude, and bearing temperature. The above data is also uploaded in real time via the industrial network.

[0073] Personnel location information is collected by equipping personnel with uniquely identified positioning cards or mining smartphones. Positioning base stations deployed in the roadway receive signals and, through time difference or signal strength analysis methods, calculate and upload the personnel's three-dimensional spatial coordinates, identity information, and movement trajectory in real time.

[0074] In S2, multi-source safety production data is fused and features are extracted. Based on equipment operating parameters and associated environmental monitoring data, a health vector representing the overall operating status of the equipment is generated by mapping time-series signals to a high-dimensional feature space. Specifically, this includes:

[0075] First, an adaptive decomposition is performed on the time-series signal formed by the operating parameters of equipment from the same monitoring target and their spatially correlated environmental monitoring data. Specifically, an initial modality count and a set of center frequency parameters containing the same number of modalities are preset as a starting point. The optimization objective is to obtain the decomposition result with the lowest envelope complexity. The envelope complexity is calculated as follows: for each modal component obtained from potential decomposition, the sum of its upper and lower envelopes is taken to form the total envelope signal, and the approximate entropy of this total envelope signal is calculated. This approximate entropy value is defined as the envelope complexity. By iteratively adjusting the modality count and center frequency parameters, the parameter combination that minimizes the envelope complexity is found, thereby adaptively determining the actual number of modes required for the final decomposition and the center frequency corresponding to each mode. Subsequently, based on these optimized parameters, variational mode decomposition is performed on the original time-series signal to obtain a series of intrinsic mode components with different center frequencies and time scales.

[0076] Secondly, for each intrinsic mode component (IMC), three features are calculated: energy entropy, sample entropy, and zero-crossing rate. The energy entropy is calculated as follows: calculate the energy value of all data points within the IMC, i.e., the square of the amplitude of each data point; then normalize these energy values ​​to a probability distribution; finally, calculate the information entropy value of this probability distribution. The sample entropy is calculated as follows: set an embedding dimension (2 in this embodiment) and a similarity tolerance threshold (0.2 times the standard deviation of the original time series signal in this embodiment); in the time series of the IMC, count the proportion of data points whose difference is less than the similarity tolerance threshold in vectors of length equal to the embedding dimension and vectors of length equal to the embedding dimension plus one; the sample entropy is the difference of the natural logarithms of these two proportions. The zero-crossing rate is calculated as follows: count the number of times the sign of adjacent data points changes in the time series of the IMC, then divide by the total length of the sequence minus 1. After the above calculations, each IMC yields a feature set containing these three feature values.

[0077] Next, the feature sets calculated from all intrinsic mode components are fused. The feature sets are arranged sequentially according to the center frequencies of each component from high to low (i.e., from small to large time scales). Before concatenation, each feature set is multiplied by a weighting coefficient, which is equal to the proportion of the energy of its corresponding intrinsic mode component (i.e., the sum of the squares of the amplitudes of all data points of that component) to the total energy of all components. The weighted feature sets are then sequentially concatenated to form a high-dimensional intermediate feature vector.

[0078] Finally, the intermediate feature vector is normalized. The method is maximum-minimum normalization, which involves iterating through all values ​​in the intermediate feature vector, finding the maximum and minimum values; for each value in the vector, subtracting the minimum value, and then dividing by the difference between the maximum and minimum values ​​(when the maximum and minimum values ​​are equal, the value is set to 0). The vector after this processing is the final output, representing the health status of the device under a specific spatiotemporal environment.

[0079] In S3, the health vector is coupled with the corresponding personnel location information and spatial topology for analysis. Through an inference process based on a probabilistic graphical structure, the probability distribution of different levels of safety risks caused by the equipment under spatial location and personnel distribution conditions is calculated, specifically including:

[0080] First, a multi-level risk propagation diagram is constructed based on the spatial topology of mine roadways. This diagram includes three types of nodes: equipment nodes, spatial region nodes, and global risk nodes. Equipment nodes represent key underground equipment (such as coal mining machines and ventilation fans), each associated with its health vector. Spatial region nodes represent continuous physical areas divided underground (such as mining faces, transport roadways, and chambers), defined by roadway connectivity and functional zoning, with each region node having a clear boundary. Global risk nodes represent the overall safety risk status of the entire mine or a specific production system. The connection relationships between nodes are as follows: equipment nodes are connected to their respective or adjacent spatial region nodes through spatial proximity, with connection weights calculated based on the spatial distance between the equipment node and the center point of the region node; spatial region nodes are connected through topological connectivity, meaning that if two regions are directly connected in the roadway network, a connection edge exists, with weights set according to roadway type, length, and ventilation conditions; each spatial region node is connected to a global risk node through influence weights, which reflect the contribution of the region's risk to the overall risk. The connection weights from equipment nodes to spatial region nodes are... The calculation uses the following formula: ;

[0081] in, Represents device node To spatial region node The connection weights, Represents device node Spatial location and spatial region nodes The Euclidean distance between the center points (unit: meters). This is a preset distance threshold (in meters), set according to the device's influence range; in this implementation, it is set to 50 meters. This formula ensures that when the distance exceeds the threshold... The time weight is zero, meaning the device only affects the neighboring area.

[0082] Secondly, the health vector is transformed into the initial risk value of the device node. For each device node, its health vector is a multi-dimensional normalized vector, which is transformed into an initial risk value through a linear mapping function. Specifically, the initial risk value... The calculation method is as follows: take the average value of all dimensions in the health vector, then subtract this average value from one, and use the result as the initial risk value. That is, if the health vector has... Each dimension has a value. The initial risk value is: ;in Indicates the first The system has several dimensions; the initial risk value ranges from 0 to 1, with higher values ​​indicating worse equipment condition and higher risk. Simultaneously, it calculates the personnel density impact factor for each spatial area node based on personnel location information. The calculation process includes three sub-steps: First, based on the physical boundaries of each spatial area node divided by spatial topology, it counts the number of personnel located within those boundaries in real time, calculating the basic static personnel density, i.e., the number of personnel divided by the area area (unit: person / square meter). Second, based on the historical sequence of personnel location information within a preset time window (e.g., the most recent 5 minutes), it calculates the frequency and direction of personnel crossing the boundaries of adjacent spatial area nodes, quantifying the intensity of personnel dynamic mobility. Mobility intensity is calculated as the sum of the total number of people entering and leaving the area within the time window, divided by the time window duration (unit: person / minute). Third, it weights and fuses the basic static personnel density and the intensity of personnel dynamic mobility, multiplying by a correction coefficient. The correction coefficient is related to the criticality level of the spatial area node in the escape route; the criticality level is set according to the necessity of the area in the preset escape route, with higher levels resulting in larger coefficients. Personnel Density Impact Factor The calculation uses the following formula: ;

[0083] here, Represents spatial region nodes The influence factor of personnel density This represents the basic static population density (normalized to the range of 0-1). This indicates the intensity of personnel mobility (normalized to the range of 0-1). This indicates the criticality level correction factor (preset value, range 1-2). and These are weighting coefficients, satisfying... ,Pick =0.6, =0.4. This factor comprehensively reflects the risk amplification effect brought about by the gathering and movement of people.

[0084] Next, the initial risk value is integrated with the personnel density influencing factor and used as the input for each spatial region node. For each spatial region node, its initial input risk value is... The initial risk value is calculated as follows: the risk value of all device nodes connected to this area. Multiply by the corresponding connection weight The sum of the results is then multiplied by the population density influence factor for that region. .

[0085] The process involves summing all device nodes connected to the node in the region. Then, multi-round probability propagation is performed through the risk propagation graph. The propagation process is based on a linear update rule: the risk probability value of each node is updated by multiplying the current risk probability values ​​of all its neighboring nodes (based on the connection edges) by their corresponding connection weights, plus its own initial input risk value (for spatial region nodes) or zero (for other nodes). Specifically, for spatial region nodes, its risk probability value... In the number of iterations Updated to: ;in It is a damping factor (taken as 0.3 in this implementation scheme). Indicates the first Each node in time The risk probability value, This indicates that all other nodes (including device nodes and adjacent region nodes) connected to this region node will be traversed. These are the connection weights, initially: ,in This represents the initial risk probability value. For a global risk node, its risk probability value is updated to the weighted sum of the risk probability values ​​of all spatial region nodes multiplied by their corresponding influence weights. The influence weights are preset based on the region area and production importance. Iteration continues until the change in the risk probability value of all nodes is less than a convergence threshold (e.g., 0.001), indicating that the probability distribution is stable.

[0086] Finally, the converged global risk node and the risk probability values ​​of each spatial region node are output as the probability distribution of different levels of safety risk. These probability values ​​range from 0 to 1 and can be directly used to assess the risk level: for example, a threshold of below 0.3 is set as low risk, 0.3-0.7 as medium risk, and above 0.7 as high risk, thus obtaining the risk level classification for each region and the entire system. The entire calculation process couples equipment status, personnel distribution, and spatial topology to quantitatively infer risk probabilities, providing a basis for subsequent scheduling decisions.

[0087] In S4, based on the probability distribution and the utility functions of different predefined scheduling strategies, an optimal gradient scheduling strategy is generated for the equipment by maximizing expected utility. The gradient scheduling strategy includes multiple ordered response levels from routine maintenance to emergency shutdown, specifically including:

[0088] First, a two-dimensional evaluation matrix is ​​constructed. The row dimension of this matrix corresponds to the risk level in the probability distribution of different levels of safety risks output in step S3. In this embodiment, it is set that if the risk level is divided into three levels: low, medium, and high, then the matrix has three rows. The column dimension of the matrix corresponds to the scheduling strategies of multiple predefined ordered response levels. These strategies are sorted from low to high response intensity, forming a continuous gradient from routine maintenance to emergency shutdown. For example, they can be specifically defined as five strategies: "record observation only," "early warning and enhanced inspection," "planned load reduction or deceleration," "area isolation and personnel evacuation," and "system-wide emergency shutdown."

[0089] Next, calculations are performed on each cell in the aforementioned evaluation matrix. Each cell is associated with a specific risk level and a specific scheduling strategy. The calculation process consists of two steps. First, the preliminary expected utility value of the cell is calculated. Specifically, the probability of occurrence of the risk level corresponding to the cell's row is obtained from the probability distribution output in step S3; the utility value generated when the risk level corresponding to the cell's column occurs is obtained. This utility value is a predefined numerical value reflecting the overall benefit (positive utility) or cost (negative utility) of implementing the strategy under that risk scenario. Multiplying the risk occurrence probability by the strategy utility value yields the preliminary expected utility value of the cell. Second, the preliminary expected utility values ​​calculated for all cells are normalized. The method is to find the maximum and minimum values ​​among all the preliminary expected utility values; for each cell's preliminary expected utility value, the minimum value is subtracted, and then divided by the difference between the maximum and minimum values. After this processing, the values ​​of all cells are scaled to between 0 and 1, which is called the normalized expected utility value.

[0090] Next, the normalized expected utility value of the cells corresponding to high-risk levels in the evaluation matrix is ​​subject to decay correction. This correction process includes four sub-steps. First, based on the current production stage (e.g., normal production, maintenance preparation, shift handover) and shift information (e.g., day shift, night shift), a dynamic risk sensitivity level is determined by querying a preset mapping table, and a corresponding baseline decay coefficient is obtained. The higher the risk sensitivity level, the smaller the baseline decay coefficient (for example, the risk sensitivity level is high during night shift production, corresponding to a baseline decay coefficient of 0.9; the risk sensitivity level is low during day shift maintenance, corresponding to a baseline decay coefficient of 1.0). Second, based on the risk level, different level correction weights are assigned to the rows of different risk levels in the evaluation matrix. The level correction weight of the row with a high risk level is greater than that of the row with a low risk level (in this implementation scheme, the weight of high-risk rows is set to 1.2, the weight of medium-risk rows is 1.0, and the weight of low-risk rows is 0.8). Third, a historical correction factor is calculated by combining historical accident frequency data of the specific spatial area (i.e., the area targeted by risk propagation in step S3) associated with the assessment matrix within the most recent preset number of days (30 days in this implementation scheme). The calculation method is as follows: count the total number of accidents that occurred in the area within the most recent preset number of days, divide by the preset number of days to obtain the daily average accident frequency; normalize the daily average accident frequency to the range of zero to one; the historical correction factor is equal to one plus the normalized daily average accident frequency (in this implementation scheme, if the normalized accident frequency is 0.15, then the historical correction factor is 1.15). This factor is positively correlated with the historical accident frequency of the same period. Fourth, the product of the above-mentioned baseline attenuation coefficient, level correction weight, and historical correction factor is used as the final dynamic attenuation factor. Using this dynamic attenuation factor, the normalized expected utility value of the corresponding high-risk level cell is multiplied and corrected, that is, the original value is multiplied by the dynamic attenuation factor, thereby attenuating its value. The values ​​of medium and low-risk level cells are usually not subject to this correction.

[0091] Finally, the optimal gradient scheduling strategy is calculated and determined. After correction, the sum of the corrected expected utility values ​​of all cells corresponding to each column (i.e., each scheduling strategy) in the evaluation matrix is ​​calculated. Specifically, for each column, the corrected values ​​of all cells in its contained rows are added together to obtain a sum representing the overall expected utility of that strategy. The sums of all columns are compared, and the scheduling strategy corresponding to the column with the highest sum is determined as the optimal gradient scheduling strategy for the device and its surrounding environment at the current moment. This strategy represents the best response level after balancing risk probability, strategy cost-effectiveness, current production sensitivity, and historical safety status.

[0092] In S5, based on the gradient scheduling strategy and GIS spatial analysis functions, specific spatial scheduling instructions are generated and executed. These instructions include control commands for related equipment, evacuation route planning for affected personnel, and production task adjustment plans, forming a closed-loop management system from risk perception to control execution. Specifically, this includes:

[0093] First, based on the probability distribution calculated in step three, a risk heat map is dynamically generated in the geographic information system, centered on the risk source equipment and reflecting the spatial spread trend of different safety risk levels. The generation process includes four steps. First, the spatial coordinates of the equipment are determined as the heat map center point. The probability values ​​of different safety risk levels in the probability distribution are converted into basic heat values ​​with different initial intensities through a preset numerical mapping relationship. The mapping relationship is: the higher the probability value, the larger the corresponding basic heat value, and the two are linearly proportional. Second, based on the spatial topological connection relationship of the mine roadways and the predetermined wind direction and velocity parameters of the ventilation network, the spatial propagation results of the above basic heat values ​​along the roadway network are calculated. The propagation calculation follows the principle of physical diffusion simulation: heat values ​​originate from the center point and diffuse outwards along connected roadways. During the diffusion process, the heat value attenuates after passing through a standard-length roadway, depending on factors such as ventilation intensity, cross-sectional size, and support conditions. The attenuation ratio is obtained by querying a preset roadway attribute table. If a roadway branches, the heat value is distributed according to the traffic capacity ratio of each branch roadway. Third, the heat value propagated to that location is dynamically enhanced by combining the personnel density influence factor calculated in step three. The enhancement method is to multiply the heat value at that location by an enhancement coefficient proportional to the personnel density influence factor at that location. The larger the personnel density influence factor, the larger the enhancement coefficient, thus reflecting the amplification effect of personnel gathering on risk perception in the heat map. Fourth, the heat value results covering all spatial locations of the entire mine after spatial propagation and personnel density enhancement processing are visualized and rendered on a geographic information system platform. During rendering, a continuous color gradient from cool to warm colors is used to fill the area based on the range of heat values. Low heat value areas are represented by blue-green tones, while high heat value areas are represented by red-yellow tones, thus generating an intuitive risk heat map. This map dynamically displays the spatial distribution and intensity of risk through color depth and range.

[0094] Secondly, based on the distribution range and color intensity of the generated risk heatmap, combined with the spatial topology of the tunnels, the key equipment sets requiring immediate intervention and the affected associated tunnel areas are identified and delineated. Specifically, in the geographic information system, a heatmap value threshold is set, and continuous geographical areas with a color depth exceeding this threshold are automatically delineated as high-risk areas. All production equipment located within these high-risk areas is analyzed and marked as the equipment set requiring immediate intervention. Simultaneously, based on tunnel connectivity, tunnel areas directly adjacent to high-risk areas and located downwind of ventilation or on essential personnel passageways are also delineated as affected associated areas and included in the overall scheduling considerations.

[0095] Next, for the identified set of critical equipment, differentiated equipment coordination control command sequences are generated based on the level of the optimal gradient scheduling strategy determined in step four. Different response levels correspond to different combinations of command content: for example, the "early warning" level only generates status alarm and recording commands; the "planned load reduction" level generates adjustment commands to reduce equipment operating power and speed; and the "emergency stop" level generates a safety interlock command to immediately cut off the power source. The command sequence must ensure the logical order and timing of actions between equipment to avoid secondary risks caused by the action of a single piece of equipment. Simultaneously, based on the risk spatial gradient (i.e., the direction and rate of color change from dark to light) shown in the risk heat map and real-time personnel location information, graded evacuation routes are planned for personnel in the affected associated roadway areas. The planning principle is to guide personnel from high-heat areas along the path with the fastest heat reduction to preset safety refuge points or main ventilation roadways. The system calculates and dynamically indicates the optimal escape path for personnel at different starting positions and provides directional guidance at major intersections.

[0096] Finally, in the simulation environment provided by the Geographic Information System (GIS), the generated equipment collaborative control command sequence and hierarchical evacuation route planning scheme were simulated and executed to verify their spatial feasibility. The simulation verification included: checking whether the control commands matched the actual spatial locations of the equipment, and whether command execution would cause spatial interference between equipment; verifying whether the planned evacuation routes were unobstructed, and whether there were any closed or blocked road sections. After successful verification, the control command sequence was sent to the corresponding field programmable logic controllers (PLCs) via the industrial control network for execution, and the evacuation route planning scheme was pushed to the mobile terminals of relevant personnel and the guidance displays within the roadways. After command execution, the system continuously collected real-time information such as equipment status feedback, personnel movement trajectories, and environmental monitoring data, and used this new data to recalculate the risk probability distribution, thereby updating the risk heat map. By comparing the evolution of the heat map before and after command execution, the actual effect of the scheduling action was evaluated, thus completing a complete closed-loop management process from risk perception, analysis and decision-making, command generation, simulation verification to on-site execution and effect feedback.

[0097] Please see Figure 2 As shown, the GIS-based intelligent integrated dispatch and management system for coal mine safety production is characterized by comprising:

[0098] The multi-source safety production data acquisition module is used to collect multi-source safety production data in coal mines. The multi-source safety production data includes environmental monitoring data, equipment operating parameters and personnel positioning information that are spatially correlated.

[0099] The equipment health status feature extraction and vectorization module integrates and extracts features from multi-source safety production data. Based on equipment operating parameters and associated environmental monitoring data, it generates a health vector representing the overall operating status of the equipment by mapping time-series signals to a high-dimensional feature space.

[0100] The risk probability inference module coupled with spatial topology performs coupled analysis with the health vector and the corresponding personnel location information and spatial topology relationship. Through the inference process based on the probability graph structure, it calculates the probability distribution of different levels of safety risks caused by the equipment under the conditions of spatial location and personnel distribution.

[0101] The gradient scheduling strategy decision optimization module generates the optimal gradient scheduling strategy for the equipment by calculating the expected utility maximization based on the probability distribution and the utility functions of different predefined scheduling strategies. The gradient scheduling strategy includes multiple ordered response levels from routine maintenance to emergency shutdown.

[0102] The GIS spatial scheduling instruction generation and closed-loop execution module, based on the gradient scheduling strategy and GIS spatial analysis function, generates and executes specific spatial scheduling instructions. The spatial scheduling instructions include control instructions for related equipment, evacuation route planning for affected personnel, and adjustment plans for production tasks, forming a closed-loop management from risk perception to control execution.

[0103] The working principle of this invention is as follows: First, it collects and spatially correlates multi-source safety production data, including underground environmental monitoring data, equipment operating parameters, and personnel positioning information. Second, it adaptively decomposes and extracts multi-dimensional features from the time-series signals of the equipment and its associated environmental parameters to generate a health vector representing the overall operating status of the equipment. Next, it couples this health vector with personnel positioning information and roadway spatial topology, constructs a multi-level risk propagation map, and performs probability inference to calculate the probability distribution of safety risks in different spatial areas. Then, based on this probability distribution and a predefined scheduling strategy utility function, it generates a gradient optimal scheduling strategy from routine maintenance to emergency shutdown through expected utility maximization calculation and dynamic correction. Finally, based on this strategy and GIS spatial analysis functions, it dynamically generates a risk heat map, identifies key intervention areas, generates and simulates equipment collaborative control commands and personnel graded evacuation channels, and finally issues and executes them, updating the risk situation using feedback data. This forms a complete management process from multi-source information perception, intelligent risk analysis, optimization strategy generation to spatial closed-loop scheduling.

[0104] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A GIS-based intelligent integrated scheduling and management method for coal mine safety production, characterized in that, Includes the following steps: S1: Collect multi-source safety production data from underground coal mines. Multi-source safety production data includes environmental monitoring data, equipment operating parameters, and personnel positioning information linked by spatial location. S2: This involves fusing and extracting features from multi-source safety production data. Based on equipment operating parameters and associated environmental monitoring data, a health vector representing the overall operating status of the equipment is generated by mapping time-series signals to a high-dimensional feature space. Specifically, this includes: Adaptive decomposition of time-series signals of equipment operating parameters and associated environmental monitoring data yields multiple intrinsic mode components with different time scales; For each intrinsic mode component, calculate the intrinsic mode component energy entropy, sample entropy and zero crossing rate to form the feature set of the intrinsic mode component; The feature sets corresponding to multiple intrinsic mode components are weighted and concatenated according to the component time scale to obtain a high-dimensional intermediate feature vector. The intermediate feature vector is normalized by its maximum and minimum values, and the processed vector is used as the health vector representing the overall operating status of the equipment. The process of obtaining the intrinsic mode components is as follows: Preset the number of initial modes and center frequency parameters required for variational mode decomposition; By introducing envelope complexity as an optimization objective, the initial number of modes and center frequency parameters are iteratively optimized to adaptively determine the number of modes required for decomposition and the center frequency of each mode. Based on the number of modes required for the optimized decomposition and the center frequency of each mode, variational mode decomposition is performed on the time-series signal to obtain multiple eigenmode components with different time scales. S3: Couple the health vector with the corresponding personnel location information and spatial topology relationship for analysis. Through the inference process based on the probabilistic graph structure, calculate the probability distribution of different levels of safety risks caused by the equipment under the conditions of spatial location and personnel distribution. S4: Based on the probability distribution and the utility functions of different predefined scheduling strategies, the optimal gradient scheduling strategy is generated for the equipment by maximizing the expected utility. The gradient scheduling strategy includes multiple ordered response levels from routine maintenance to emergency shutdown. S5: Based on the gradient scheduling strategy and GIS spatial analysis function, it generates and executes specific spatial scheduling instructions, including control instructions for related equipment, evacuation route planning for affected personnel, and adjustment plans for production tasks, forming a closed-loop management from risk perception to control execution.

2. The GIS-based intelligent integrated scheduling and management method for coal mine safety production according to claim 1, characterized in that, S3 specifically includes: Based on spatial topology, a multi-level risk propagation graph is constructed, including device nodes, spatial region nodes, and global risk nodes. Device nodes and spatial region nodes are connected through spatial proximity, spatial region nodes are connected through topological connectivity, and spatial region nodes and global risk nodes are connected through influence weights. The health vector is converted into the initial risk value of the device node, and the personnel density influence factor of each spatial area node is calculated based on the personnel location information. The initial risk value is integrated with the personnel density influencing factor and used as the input for each spatial region node. The probability propagation is carried out through multiple rounds of iterations through the risk propagation map until the risk probability value of each node converges. Output the risk probability values ​​of global risk nodes and nodes in each spatial region as the probability distribution of different levels of security risks.

3. The GIS-based intelligent integrated scheduling and management method for coal mine safety production according to claim 2, characterized in that, The calculation process for the personnel density influence factor is as follows: The physical boundaries of each spatial region node are determined based on the spatial topology, and the number of people located within the physical boundaries in real time is counted to calculate the basic static personnel density of the spatial region. Based on the historical sequence of personnel location information within a preset time window, the frequency and direction of personnel crossing the boundaries of adjacent spatial area nodes are calculated to quantify the intensity of personnel dynamic mobility in the spatial area. The basic static population density and the intensity of dynamic population mobility are weighted and fused together, and then multiplied by a correction coefficient related to the criticality level of spatial area nodes in the escape route to obtain the population density influence factor.

4. The GIS-based intelligent integrated scheduling and management method for coal mine safety production according to claim 1, characterized in that, S4 specifically includes: Construct a two-dimensional evaluation matrix. The row dimension of the two-dimensional evaluation matrix corresponds to different risk levels in the probability distribution, and the column dimension corresponds to the scheduling strategies of multiple predefined ordered response levels. For each cell in the evaluation matrix, the product of the probability of occurrence of the risk level and the utility value of the corresponding scheduling strategy under the risk level is calculated to obtain the preliminary expected utility value, and the preliminary expected utility values ​​of all cells are normalized. The normalized expected utility value of the cells corresponding to high-risk levels in the evaluation matrix is ​​subject to decay correction; Calculate the sum of the corrected expected utility values ​​for each column of scheduling policies, and determine the scheduling policy corresponding to the column with the highest sum as the optimal gradient scheduling policy at the current time.

5. The GIS-based intelligent integrated scheduling and management method for coal mine safety production according to claim 4, characterized in that, The process of attenuating the normalized expected utility value of the cells corresponding to high-risk levels in the evaluation matrix includes: Based on the current production stage and shift information, determine the dynamic risk sensitivity level and obtain the corresponding baseline attenuation coefficient accordingly. Based on the level of risk, different level adjustment weights are assigned to the rows of different risk levels in the assessment matrix; Based on the historical accident frequency data of the corresponding spatial area within the most recent preset number of days, a historical correction factor is calculated. The historical correction factor is positively correlated with the accident frequency of the same historical period. The product of the baseline decay factor, the grade correction weight, and the historical correction factor is used as the final dynamic decay factor to multiply and correct the normalized expected utility value of the corresponding cell.

6. The GIS-based intelligent integrated scheduling and management method for coal mine safety production according to claim 1, characterized in that, S5 specifically includes: Based on probability distribution, a risk heat map is dynamically generated in GIS, centered on the equipment, reflecting the spatial spread trend of different safety risk levels; Based on the distribution range and intensity of the risk heat map, combined with spatial topological relationships, the equipment sets requiring immediate intervention and the affected associated roadway areas are identified and delineated. For the equipment set, based on the level of the gradient scheduling strategy, a differentiated sequence of equipment collaborative control instructions is generated. At the same time, based on the risk gradient of the risk heat map and personnel location information, a graded evacuation channel is planned for personnel in the associated roadway area. In GIS, simulate the sequence of coordinated control commands for equipment and hierarchical evacuation routes. After verifying spatial feasibility, issue the commands for execution and update the risk heat map based on real-time data after execution to complete closed-loop verification.

7. The GIS-based intelligent integrated scheduling and management method for coal mine safety production according to claim 6, characterized in that, The process of generating the risk heatmap is as follows: Using the spatial location of the equipment as the thermal center point, the probability values ​​of different safety risk levels in the probability distribution are mapped to basic thermal values ​​with different initial intensities. Based on the orientation of the tunnel connections and ventilation network in the spatial topology, the spatial propagation of the basic thermal value along the tunnel network is calculated. By incorporating the influence factor of population density, the thermal values ​​propagated to various spatial locations are dynamically enhanced; The heat values, after propagation and enhancement, are rendered in GIS as a risk heat map with a continuous color gradient.

8. A GIS-based intelligent integrated dispatch and management system for coal mine safety production, characterized in that: The method for implementing the GIS-based intelligent integrated scheduling and management method for coal mine safety production according to any one of claims 1-7 includes: The multi-source safety production data acquisition module is used to collect multi-source safety production data in coal mines. The multi-source safety production data includes environmental monitoring data, equipment operating parameters and personnel positioning information that are spatially correlated. The equipment health status feature extraction and vectorization module integrates and extracts features from multi-source safety production data. Based on equipment operating parameters and associated environmental monitoring data, it generates a health vector representing the overall operating status of the equipment by mapping time-series signals to a high-dimensional feature space. The risk probability inference module coupled with spatial topology performs coupled analysis with the health vector and the corresponding personnel location information and spatial topology relationship. Through the inference process based on the probability graph structure, it calculates the probability distribution of different levels of safety risks caused by the equipment under the conditions of spatial location and personnel distribution. The gradient scheduling strategy decision optimization module generates the optimal gradient scheduling strategy for the equipment by calculating the expected utility maximization based on the probability distribution and the utility functions of different predefined scheduling strategies. The gradient scheduling strategy includes multiple ordered response levels from routine maintenance to emergency shutdown. The GIS spatial scheduling instruction generation and closed-loop execution module, based on the gradient scheduling strategy and GIS spatial analysis function, generates and executes specific spatial scheduling instructions. The spatial scheduling instructions include control instructions for related equipment, evacuation route planning for affected personnel, and adjustment plans for production tasks, forming a closed-loop management from risk perception to control execution.

Images