Safety intelligent identification and early warning system for non-coal mine mountain top plate

Abstract

The invention discloses a safety intelligent identification and early warning system for a non-coal mine mountain top plate, and the system specifically comprises a multi-source sensor data collection module, a data processing and feature extraction module, an intelligent identification and risk assessment module, a grading early warning and decision support module, and a linkage control and emergency disposal module. According to the invention, the early warning response is carried out according to the score corresponding to the specific risk index, the multi-level and multi-dimensional early warning information is generated, and the targeted disposal suggestions are provided according to the early warning information, so that corresponding decision support is conveniently carried out according to the response levels of different risks, and the risk assessment efficiency is improved. And linkage control is performed on the early warning information and the mining equipment through the linkage control and emergency disposal module, so that the risk response and the equipment can be subjected to linkage control through linkage of different types of control measures in the process that the risk level is in a general risk, a relatively large risk and a major risk, and the risk level is improved. The intervention measures are actively triggered in different risk scenes, and the accident loss is reduced.

Description

Technical Field

[0001] This invention relates to the field of safety identification and early warning technology, specifically to an intelligent identification and early warning system for the safety of roof slabs in non-coal mines. Background Technology

[0002] The intelligent identification and early warning system for roof safety in non-coal mines is an important technical means to ensure safe production in mines. As a high-risk industry, roof safety management has always been a core aspect of safe production in non-coal mines. In the process of non-coal mining, the roof refers to the rock strata covering the ore body above the mine. The safety condition of the roof is directly related to the safety of mining operations. By using artificial intelligence technology, especially image recognition and big data analysis technology, the system can monitor and warn of potential safety hazards on the roof during non-coal mining. When a safety hazard is detected, an alarm message will be sent to relevant personnel immediately so that timely measures can be taken to deal with it, thereby improving the safety of mining operations and reducing the accident rate.

[0003] Currently, in the process of safety identification and early warning of roof slabs in non-coal mines, different safety hazards lead to different accident outcomes. Therefore, it is impossible to differentiate and issue warnings based on the actual severity of the safety hazards during the identification and early warning process. This results in the inability to respond to the warning information in a timely manner and provide targeted decision support, leading to a lack of effective synchronization and coordination between the response results and the warning information. This can easily lead to a waste of resources and an expansion of losses. Summary of the Invention

[0004] This invention provides an intelligent identification and early warning system for safety on the roof of non-coal mines. It effectively addresses the problems mentioned in the background section regarding the current safety identification and early warning process for roofs in non-coal mines. Because different safety hazards lead to different accident outcomes, the system cannot differentiate between hazards based on their actual severity during the identification and early warning process. This results in a failure to respond promptly to the early warning information, hindering timely decision support and leading to ineffective synchronization between the response and the warning information. Consequently, this system is prone to resource waste and increased losses.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent identification and early warning system for roof safety in non-coal mines, which accurately and efficiently ensures roof safety in non-coal mines through intelligent identification and early warning processing, and effectively prevents and mitigates the harm caused by production accidents and disasters;

[0006] Specifically, it includes a multi-source sensor data acquisition module, a data processing and feature extraction module, an intelligent identification and risk assessment module, a graded early warning and decision support module, and a linkage control and emergency response module;

[0007] The multi-source sensor data acquisition module collects physical, mechanical, and state data of the roof slab in non-coal mines in real time by deploying various sensors and intelligent monitoring equipment. The data processing and feature extraction module is designed to process the input raw data and accurately extract features that reflect the stability of the roof.

[0008] The intelligent identification and risk assessment module identifies and judges the current state of the roof based on multi-source feature data, and quantifies the specific risk level according to the roof state. The graded early warning and decision support module generates multi-level and multi-dimensional early warning information based on the quantitative risk assessment results. The linkage control and emergency response module links the early warning information with the mining equipment to facilitate proactive intervention measures and reduce accident losses.

[0009] According to the above technical solution, the multi-source sensor data acquisition module includes a micro-vibration monitoring unit, a pressure monitoring unit, a displacement monitoring unit, and an environmental monitoring unit;

[0010] The micro-vibration monitoring unit collects micro-vibration signals generated by the fracturing of the roof rock layer by using a high-sensitivity micro-vibration sensor.

[0011] The pressure monitoring unit uses pressure sensors to monitor the stress changes inside the rock strata and the stress changes of the support structure in real time, and identifies stress concentration areas to determine the effectiveness of the support and the stability of the rock strata.

[0012] The displacement monitoring unit uses a displacement sensor, a laser rangefinder, and a tilt sensor to monitor displacement and tilt data.

[0013] Specifically, displacement sensors and laser rangefinders are used to monitor the rate and cumulative amount of subsidence of the roof, while tilt sensors are used to monitor the tilt angle of the roof in order to assess the risk of roof collapse.

[0014] The environmental monitoring unit uses temperature and humidity sensors, gas sensors, and dust concentration sensors to monitor environmental temperature and humidity, gas concentration, and dust concentration in real time, respectively. By monitoring environmental parameter data, it determines the impact of environmental disturbances on the stability of the roof.

[0015] According to the above technical solution, in the specific arrangement of the micro-vibration monitoring unit, pressure monitoring unit, displacement monitoring unit and environmental monitoring unit, the multi-source sensor data acquisition module uses virtual technology to simulate a three-dimensional model of the mine roadway, and marks the specific deployment location of the monitoring sensor equipment of each monitoring unit in the three-dimensional model by annotation, and uses different marking colors to distinguish and mark different types of sensors.

[0016] Among them, micro-vibration sensors are arranged in key areas of the roadway roof, pressure sensors are installed at the free ends of anchor bolts and anchor cables, displacement sensors and laser rangefinders are deployed in key positions of the roof, and tilt sensors are installed on the roof support structure.

[0017] Temperature and humidity sensors, gas sensors, and dust concentration sensors are installed in a distributed manner at multiple monitoring points in the mine roadway.

[0018] According to the above technical solution, the data processing and feature extraction module refers to performing preliminary processing on the collected multi-source data information to ensure the accuracy of the data, while extracting key features.

[0019] In the data processing process, the data is cleaned and standardized, high-frequency noise from multi-source sensor data is filtered out, abnormal data is deleted and missing data is filled in, and the data format is unified after data processing.

[0020] When extracting key features, the main approach is to use preset rules to support rapid subsequent judgment. Specifically, this includes extracting key features from micro-vibration monitoring data, pressure monitoring data, displacement monitoring data, and environmental monitoring data.

[0021] According to the above technical solution, the intelligent identification and risk assessment module is based on multi-source feature data to comprehensively identify and judge the current state of the roof and conduct risk assessment. Specifically, it includes a multi-model fusion identification unit and a risk quantification unit.

[0022] The multi-model fusion recognition unit includes a traditional threshold model and a machine learning model, both of which are built into the central server.

[0023] Traditional threshold models incorporate thresholds calibrated from historical mine data, setting displacement rate thresholds and stress exceedance thresholds. Machine learning models employ LSTM neural networks, taking into account the roof subsidence, displacement rate, and support stress data from the previous 24 hours, and outputting predicted displacement values ​​for the next 2 hours to predict trends.

[0024] The risk quantification unit uses DS evidence theory to fuse the identification results of multiple models to output a comprehensive risk index. Specifically, it uses traditional threshold model and machine learning model as the basic probability allocation to evaluate and calculate the comprehensive risk index.

[0025] The comprehensive risk index is expressed as a score, specifically from 0 to 10. Among them, 0-3 points indicate low risk, 3-7 points indicate moderate risk, 7-10 points indicate significant risk, and 10 points indicate major risk.

[0026] According to the above technical solution, in the specific risk assessment process, the intelligent identification and risk assessment module represents specific risk warning information in the form of a risk heat map in a three-dimensional model of a mine roadway simulated by virtual technology. It colors and marks these areas according to risk indices to visually display the risk level of each risk area, as detailed below:

[0027] At a risk index of 0-3, a blue marker indicates a low risk; at 3-7, a yellow marker indicates a moderate risk; at 7-9, an orange marker indicates a significant risk; and at a risk index of 10, a red marker indicates a major risk.

[0028] Furthermore, in the virtual 3D model, when the risk warning is at the level of general risk, major risk, or significant risk, the specific location of the risk area needs to be marked and represented in the virtual 3D model according to coordinates.

[0029] According to the above technical solution, the graded early warning and decision support module assesses specific risks as low risk, general risk, relatively high risk, and major risk based on the risk index. During the specific early warning process, it needs to respond to the early warning according to the score corresponding to the specific risk index, and at the same time provide corresponding decision support according to the response level of different risks.

[0030] According to the above technical solution, the hierarchical early warning and decision support module, in specific hierarchical early warning scenarios, includes the following situations:

[0031] When the risk index score is 0-3, under low-risk conditions, the on-duty personnel will be notified via platform pop-up and SMS, prompting on-site personnel to strengthen patrols;

[0032] When the risk index score is 3-7, under normal risk conditions, the on-site sound and light alarm device is activated to promptly warn on-site personnel through buzzers and flashing LED lights in the on-site roadway, and to notify on-site workers to evacuate to the safe area.

[0033] When the risk index score is 7-9, under relatively high risk conditions, the support equipment is triggered to automatically replenish pressure, increase the stroke of the hydraulic support, and automatically cut off the power supply to the dangerous area.

[0034] When the risk index score is 10, in the event of a major risk, an emergency evacuation procedure needs to be initiated, evacuation instructions should be played repeatedly using on-site broadcasting equipment, and the matter should be reported to the superior regulatory authorities.

[0035] According to the above technical solution, the linkage control and emergency response module is based on hierarchical early warning and decision support. In the process of early warning according to risk level, it deeply integrates risk early warning with mining equipment to realize linkage control under different risk scenarios and reduce losses by automatically executing linkage control measures.

[0036] During periods of general risk, significant risk, and major risk, specific coordinated control measures include support equipment, ventilation equipment, positioning equipment, and broadcasting equipment.

[0037] According to the above technical solution, the linkage control and emergency response module performs the following specific functions during linkage control and emergency response:

[0038] Linking the support equipment with the hydraulic support control system, the system automatically increases the support force of the support when general risk, major risk and critical risk warnings are triggered, so as to enhance the support strength of the roof.

[0039] When general risk, major risk and critical risk warnings are triggered, roof collapse will cause gas accumulation. The working frequency of on-site ventilation equipment can be adjusted through linkage control to increase the gas dilution rate by increasing air volume.

[0040] When general risk, major risk and critical risk warnings are triggered, on-site personnel are located using positioning devices, and their real-time location within the danger zone is obtained. The optimal evacuation route is planned based on the personnel location and the danger zone, and the planned optimal evacuation route is sent to on-site personnel via SMS and APP push.

[0041] Broadcasting equipment refers to equipment that, when a general risk, a relatively large risk, or a major risk warning is triggered, plays pre-recorded warning messages to alert on-site personnel and responds in conjunction with on-site LED screens to display warning signs and warning messages.

[0042] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0043] 1. The graded early warning and decision support module facilitates early warning responses based on the quantitative assessment results of risks and the scores corresponding to specific risk indices. It generates multi-level and multi-dimensional early warning information and provides targeted handling suggestions based on the early warning information. This allows for corresponding decision support based on different risk response levels. By responding to different early warning decisions under different risk index conditions, graded early warnings are achieved according to the actual situation, enabling rapid judgment of the current risk status based on early warning information. The linkage control and emergency response module links early warning information with mining equipment. During the stages of general risk, significant risk, and major risk, different types of control measures are linked to control the risk response and equipment, enabling proactive intervention measures to be triggered in different risk scenarios and reducing accident losses.

[0044] 2. By deploying various sensors and intelligent monitoring equipment through a multi-source sensor data acquisition module, real-time data on the physical, mechanical, and state conditions of the roof slab in non-coal mines are collected. This enables comprehensive monitoring of the roof's micro-vibrations, pressure, displacement, and environment, providing accurate raw data support for subsequent analysis. Simultaneously, virtual technology is used to simulate a three-dimensional model of the mine roadway. Within the three-dimensional model, the specific deployment locations of each monitoring sensor are marked using annotations. Different types of sensors are distinguished by color, facilitating rapid identification of sensor locations and types at different nodes, and enabling timely deployment and maintenance.

[0045] 3. The data processing and feature extraction module can process the input raw data. By cleaning and standardizing the data, the accuracy of the data is ensured. Based on preset rules, features reflecting the stability of the roof are accurately extracted, providing accurate data input for subsequent intelligent identification and early warning. Based on the processed and feature-extracted multi-source feature data, combined with the intelligent identification and risk assessment module, the current state of the roof is identified and judged. The specific risk level is quantified according to the roof state. By using multi-model fusion identification and risk quantification processing, traditional threshold models and machine learning models are used as basic probability allocations to evaluate and calculate the comprehensive risk index, thereby realizing multi-level risk assessment. Attached Figure Description

[0046] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0047] In the attached diagram:

[0048] Figure 1 This is an architecture diagram of the intelligent identification and early warning system of the present invention. Detailed Implementation

[0049] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0050] Example: Figure 1 As shown, the present invention provides a technical solution: an intelligent identification and early warning system for roof safety in non-coal mines. Through intelligent identification and early warning processing, it can accurately and efficiently ensure the roof safety of non-coal mines and effectively prevent and mitigate the harm caused by production accidents and disasters.

[0051] Specifically, it includes a multi-source sensor data acquisition module, a data processing and feature extraction module, an intelligent identification and risk assessment module, a graded early warning and decision support module, and a linkage control and emergency response module;

[0052] The multi-source sensor data acquisition module collects real-time data on the physical, mechanical, and state conditions of the roof slab in non-coal mines by deploying various sensors and intelligent monitoring equipment. The data processing and feature extraction module is designed to process the input raw data and accurately extract features that reflect the stability of the roof.

[0053] The intelligent identification and risk assessment module identifies and judges the current state of the roof based on multi-source feature data, and quantifies the specific risk level according to the roof state. The graded early warning and decision support module generates multi-level and multi-dimensional early warning information based on the quantitative risk assessment results. The linkage control and emergency response module links the early warning information with mining equipment to facilitate proactive intervention measures and reduce accident losses.

[0054] Based on the above technical solution, the multi-source sensor data acquisition module includes a micro-vibration monitoring unit, a pressure monitoring unit, a displacement monitoring unit, and an environmental monitoring unit;

[0055] The micro-vibration monitoring unit uses a high-sensitivity micro-vibration sensor to collect micro-vibration signals generated by the fracturing of the roof rock layer;

[0056] The pressure monitoring unit uses pressure sensors to monitor the stress changes inside the rock strata and the stress changes of the support structure in real time, and identifies stress concentration areas to determine the effectiveness of the support and the stability of the rock strata.

[0057] The displacement monitoring unit uses displacement sensors, laser rangefinders, and tilt sensors to monitor displacement and tilt data.

[0058] Specifically, displacement sensors and laser rangefinders are used to monitor the rate and cumulative amount of subsidence of the roof, while tilt sensors are used to monitor the tilt angle of the roof in order to assess the risk of roof collapse.

[0059] The environmental monitoring unit uses temperature and humidity sensors, gas sensors, and dust concentration sensors to monitor the ambient temperature and humidity, gas concentration, and dust concentration in real time, respectively. By monitoring environmental parameter data, it determines the impact of environmental disturbances on the stability of the roof.

[0060] Based on the above technical solution, in the specific arrangement of the micro-vibration monitoring unit, pressure monitoring unit, displacement monitoring unit and environmental monitoring unit, the multi-source sensor data acquisition module uses virtual technology to simulate a three-dimensional model of the mine roadway. In the three-dimensional model, the specific deployment location of the monitoring sensor equipment of each monitoring unit is marked by annotation. At the same time, different marking colors are used to distinguish and mark different types of sensors.

[0061] Among them, micro-vibration sensors are arranged in key areas of the roadway roof, pressure sensors are installed at the free ends of anchor bolts and anchor cables, displacement sensors and laser rangefinders are deployed in key positions of the roof, and tilt sensors are installed on the roof support structure.

[0062] Temperature and humidity sensors, gas sensors, and dust concentration sensors are installed in a distributed manner at multiple monitoring points in the mine roadway.

[0063] Based on the above technical solution, the data processing and feature extraction module refers to the preliminary processing of collected multi-source data information to ensure the accuracy of the data, while extracting key features.

[0064] In the data processing process, the data is cleaned and standardized, high-frequency noise from multi-source sensor data is filtered out, abnormal data is deleted and missing data is filled in, and the data format is unified after data processing.

[0065] When extracting key features, the main approach is to use preset rules to support rapid subsequent judgment. Specifically, this includes extracting key features from micro-vibration monitoring data, pressure monitoring data, displacement monitoring data, and environmental monitoring data.

[0066] Based on the above technical solution, the intelligent identification and risk assessment module is based on multi-source feature data to comprehensively identify and judge the current state of the roof and conduct risk assessment. Specifically, it includes a multi-model fusion identification unit and a risk quantification unit.

[0067] The multi-model fusion recognition unit includes a traditional threshold model and a machine learning model, both of which are built into the central server.

[0068] Traditional threshold models incorporate thresholds calibrated from historical mine data, setting displacement rate thresholds and stress exceedance thresholds. Machine learning models employ LSTM neural networks, taking into account the roof subsidence, displacement rate, and support stress data from the previous 24 hours, and outputting predicted displacement values ​​for the next 2 hours to predict trends.

[0069] The risk quantification unit uses DS evidence theory to fuse the identification results of multiple models to output a comprehensive risk index. Specifically, it uses the traditional threshold model and the machine learning model as the basic probability allocation (BPA) to evaluate and calculate the comprehensive risk index.

[0070] The comprehensive risk index is expressed as a score, specifically from 0 to 10. Among them, 0-3 points indicate low risk, 3-7 points indicate moderate risk, 7-10 points indicate significant risk, and 10 points indicate major risk.

[0071] Based on the above technical solution, the intelligent identification and risk assessment module, during the specific risk assessment process, represents specific risk warning information in the form of a risk heat map within a 3D model of a mine roadway simulated by virtual technology. It uses color coding according to risk indices to visually display the risk level of each risk area, as detailed below:

[0072] At a risk index of 0-3, a blue marker indicates a low risk; at 3-7, a yellow marker indicates a moderate risk; at 7-9, an orange marker indicates a significant risk; and at a risk index of 10, a red marker indicates a major risk.

[0073] Furthermore, in the virtual 3D model, when the risk warning is at the level of general risk, major risk, or significant risk, the specific location of the risk area needs to be marked and represented in the virtual 3D model according to coordinates.

[0074] Based on the above technical solution, the graded early warning and decision support module assesses specific risks as low risk, general risk, relatively high risk, and major risk according to the risk index. During the specific early warning process, it needs to respond to the early warning according to the score corresponding to the specific risk index, and at the same time provide corresponding decision support according to the response level of different risks.

[0075] Based on the above technical solution, the tiered early warning and decision support module includes the following specific situations when issuing tiered early warnings:

[0076] When the risk index score is 0-3, under low-risk conditions, the on-duty personnel will be notified via platform pop-up and SMS, prompting on-site personnel to strengthen patrols;

[0077] When the risk index score is 3-7, under normal risk conditions, the on-site sound and light alarm device is activated to promptly warn on-site personnel through buzzers and flashing LED lights in the on-site roadway, and to notify on-site workers to evacuate to the safe area.

[0078] When the risk index score is 7-9, under relatively high risk conditions, the support equipment is triggered to automatically replenish pressure, increase the stroke of the hydraulic support, and automatically cut off the power supply to the dangerous area.

[0079] When the risk index score is 10, in the event of a major risk, an emergency evacuation procedure needs to be initiated, evacuation instructions should be played repeatedly using on-site broadcasting equipment, and the matter should be reported to the superior regulatory authorities.

[0080] Based on the above technical solution, the linkage control and emergency response module is based on hierarchical early warning and decision support. In the process of early warning according to risk level, it deeply integrates risk early warning with mining equipment to realize linkage control under different risk scenarios and reduce losses by automatically executing linkage control measures.

[0081] During periods of general risk, significant risk, and major risk, specific coordinated control measures include support equipment, ventilation equipment, positioning equipment, and broadcasting equipment.

[0082] Based on the above technical solution, the linkage control and emergency response module performs the following specific functions during linkage control and emergency response:

[0083] The support equipment is linked with the hydraulic support control system. When a general risk, a major risk, or a critical risk warning is triggered, the support force of the support is automatically increased, specifically from 20MPa to 40MPa, to enhance the support strength of the roof.

[0084] When the warnings for general risk, major risk and critical risk are triggered, the collapse of the roof will cause gas accumulation. The working frequency of the on-site ventilation equipment will be adjusted through linkage control. Specifically, the power of the ventilation equipment will be increased from 40Hz to 60Hz to increase the gas dilution rate by increasing the air volume.

[0085] When general risk, major risk and critical risk warnings are triggered, on-site personnel are located using positioning devices, and their real-time location within the danger zone is obtained. The optimal evacuation route is planned based on the personnel location and the danger zone, and the planned optimal evacuation route is sent to on-site personnel via SMS and APP push.

[0086] Broadcasting equipment refers to equipment that, when a general risk, a relatively large risk, or a major risk warning is triggered, plays pre-recorded warning messages to alert on-site personnel and responds in conjunction with on-site LED screens to display warning signs and warning messages.

[0087] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A smart identification and early warning system for roof safety in non-coal mines, characterized in that: Intelligent identification and early warning processing can accurately and efficiently ensure the safety of the roof in non-coal mines, and effectively prevent and mitigate the harm caused by production accidents and disasters. Specifically, it includes a multi-source sensor data acquisition module, a data processing and feature extraction module, an intelligent identification and risk assessment module, a graded early warning and decision support module, and a linkage control and emergency response module; The multi-source sensor data acquisition module collects physical, mechanical, and state data of the roof slab in non-coal mines in real time by deploying various sensors and intelligent monitoring equipment. The data processing and feature extraction module is designed to process the input raw data and accurately extract features that reflect the stability of the roof. The intelligent identification and risk assessment module identifies and judges the current state of the roof based on multi-source feature data, and quantifies the specific risk level according to the roof state. The graded early warning and decision support module generates multi-level and multi-dimensional early warning information based on the quantitative risk assessment results. The linkage control and emergency response module links the early warning information with the mining equipment for linkage control.

2. The intelligent identification and early warning system for roof safety in non-coal mines according to claim 1, characterized in that: The multi-source sensor data acquisition module includes a micro-vibration monitoring unit, a pressure monitoring unit, a displacement monitoring unit, and an environmental monitoring unit; The micro-vibration monitoring unit collects micro-vibration signals generated by the fracturing of the roof rock layer by using a high-sensitivity micro-vibration sensor. The pressure monitoring unit uses pressure sensors to monitor changes in internal stress of the rock strata and stress changes in the support structure in real time, and identifies areas of stress concentration. The displacement monitoring unit uses a displacement sensor, a laser rangefinder, and a tilt sensor to monitor displacement and tilt data. Specifically, displacement sensors and laser rangefinders are used to monitor the rate and cumulative amount of subsidence of the roof, while tilt sensors are used to monitor the tilt angle of the roof in order to assess the risk of roof collapse. The environmental monitoring unit uses temperature and humidity sensors, gas sensors, and dust concentration sensors to monitor environmental temperature and humidity, gas concentration, and dust concentration in real time, respectively. By monitoring environmental parameter data, it determines the impact of environmental disturbances on the stability of the roof.

3. The intelligent identification and early warning system for roof safety in non-coal mines according to claim 1, characterized in that: In the specific arrangement of the micro-vibration monitoring unit, pressure monitoring unit, displacement monitoring unit and environmental monitoring unit, the multi-source sensor data acquisition module uses virtual technology to simulate a three-dimensional model of the mine roadway. Among them, micro-vibration sensors are arranged in key areas of the roadway roof, pressure sensors are installed at the free ends of anchor bolts and anchor cables, displacement sensors and laser rangefinders are deployed in key positions of the roof, and tilt sensors are installed on the roof support structure. Temperature and humidity sensors, gas sensors, and dust concentration sensors are installed in a distributed manner at multiple monitoring points in the mine roadway.

4. The intelligent identification and early warning system for roof safety in non-coal mines according to claim 1, characterized in that: The data processing and feature extraction module refers to the preliminary processing of collected multi-source data information, while extracting key features; In the data processing process, the data is cleaned and standardized, high-frequency noise from multi-source sensor data is filtered out, abnormal data is deleted and missing data is filled in, and the data format is unified after data processing. When extracting key features, feature extraction is performed based on preset rules to support subsequent rapid judgment. Specifically, this includes extracting key features from micro-vibration monitoring data, pressure monitoring data, displacement monitoring data, and environmental monitoring data.

5. The intelligent identification and early warning system for roof safety in non-coal mines according to claim 1, characterized in that: The intelligent identification and risk assessment module is based on multi-source feature data to comprehensively identify and judge the current state of the roof and conduct risk assessment. Specifically, it includes a multi-model fusion identification unit and a risk quantification unit. The multi-model fusion recognition unit includes a traditional threshold model and a machine learning model, both of which are built into the central server. Traditional threshold models incorporate thresholds calibrated from historical mine data, setting displacement rate thresholds and stress exceedance thresholds. Machine learning models employ LSTM neural networks, taking into account the roof subsidence, displacement rate, and support stress data from the previous 24 hours, and outputting predicted displacement values ​​for the next 2 hours to predict trends. The risk quantification unit uses DS evidence theory to fuse the identification results of multiple models to output a comprehensive risk index. Specifically, it uses traditional threshold model and machine learning model as the basic probability allocation to evaluate and calculate the comprehensive risk index. The comprehensive risk index is expressed as a score, specifically from 0 to 10. Among them, 0-3 points indicate low risk, 3-7 points indicate moderate risk, 7-10 points indicate significant risk, and 10 points indicate major risk.

6. The intelligent identification and early warning system for roof safety in non-coal mines according to claim 5, characterized in that: In the specific risk assessment process, the intelligent identification and risk assessment module represents specific risk warning information in the form of a risk heat map in a three-dimensional model of a mine roadway simulated by virtual technology. It uses color coding according to risk indices to visually display the risk level of each risk area, as detailed below: At a risk index of 0-3, a blue marker indicates a low risk; at 3-7, a yellow marker indicates a moderate risk; at 7-9, an orange marker indicates a significant risk; and at a risk index of 10, a red marker indicates a major risk. Furthermore, in the virtual 3D model, when the risk warning is at the level of general risk, major risk, or significant risk, the specific location of the risk area needs to be marked and represented in the virtual 3D model according to coordinates.

7. A smart identification and early warning system for roof safety in non-coal mines according to claim 5, characterized in that: The tiered early warning and decision support module assesses specific risks as low risk, general risk, relatively high risk, and major risk based on risk indices. During the early warning process, it responds to the early warning based on the score corresponding to the specific risk index and provides corresponding decision support based on the response level of different risks.

8. A smart identification and early warning system for roof safety in non-coal mines according to claim 7, characterized in that: The tiered early warning and decision support module, when issuing specific tiered early warnings, includes the following situations: When the risk index score is 0-3, under low-risk conditions, the on-duty personnel will be notified via platform pop-up and SMS, prompting on-site personnel to strengthen patrols; When the risk index score is 3-7, under normal risk conditions, the on-site sound and light alarm device is activated to promptly warn on-site personnel through buzzers and flashing LED lights in the on-site roadway, and to notify on-site workers to evacuate to the safe area. When the risk index score is 7-9, under relatively high risk conditions, the support equipment is triggered to automatically replenish pressure, increase the stroke of the hydraulic support, and automatically cut off the power supply to the dangerous area. When the risk index score is 10, in the event of a major risk, an emergency evacuation procedure needs to be initiated, evacuation instructions should be played repeatedly using on-site broadcasting equipment, and the matter should be reported to the superior regulatory authorities.

9. A smart identification and early warning system for roof safety in non-coal mines according to claim 8, characterized in that: The linkage control and emergency response module is based on hierarchical early warning and decision support. In the process of early warning according to risk level, it deeply integrates risk early warning with mining equipment to realize linkage control under different risk scenarios and reduce losses by automatically executing linkage control measures. During periods of general risk, significant risk, and major risk, specific coordinated control measures include support equipment, ventilation equipment, positioning equipment, and broadcasting equipment.

10. A smart identification and early warning system for roof safety in non-coal mines according to claim 9, characterized in that: The linkage control and emergency response module performs the following specific actions during linkage control and emergency response: Linking the support equipment with the hydraulic support control system, the system automatically increases the support force of the support when general risk, major risk and critical risk warnings are triggered, so as to enhance the support strength of the roof. When general risk, major risk and critical risk warnings are triggered, roof collapse will cause gas accumulation. The working frequency of on-site ventilation equipment can be adjusted through linkage control to increase the gas dilution rate by increasing air volume. When general risk, major risk and critical risk warnings are triggered, on-site personnel are located using positioning devices, and their real-time location within the danger zone is obtained. The optimal evacuation route is planned based on the personnel location and the danger zone, and the planned optimal evacuation route is sent to on-site personnel via SMS and APP push. Broadcasting equipment refers to equipment that, when a general risk, a relatively large risk, or a major risk warning is triggered, plays pre-recorded warning messages to alert on-site personnel and responds in conjunction with on-site LED screens to display warning signs and warning messages.

Images