Mine unmanned intelligent management and control system

Abstract

This invention discloses an unmanned intelligent management and control system for mines. The system includes a data acquisition layer, an edge computing layer, a cloud-based intelligent management and control platform, an execution and control layer, and a human-machine interaction terminal. By constructing a collaborative architecture of cloud, edge, and terminal, it achieves unified acquisition of all mine data, edge preprocessing, and deep cloud-based integration. Based on an intelligent analysis engine and a 3D visualization engine, it forms a comprehensive intelligent management and control capability integrating predictive equipment maintenance, intelligent ventilation control, multi-system emergency linkage, and closed-loop management of safety risks. This invention breaks down the information and control barriers between traditional subsystems, elevating discrete monitoring and alarms to autonomous and collaborative intelligent decision-making and closed-loop control, significantly improving the reliability of mine equipment, production continuity, emergency response speed, and safety level.

Description

Technical Field

[0001] This invention relates to the field of intelligent mine and safe production management technology, and in particular to an unmanned intelligent management system for mines. Background Technology

[0002] With the development of technologies such as the Internet of Things, big data, and artificial intelligence, intelligent construction of mines has become a specific direction for the industry's transformation and upgrading. Relevant national departments have successively issued guiding opinions to promote the deep integration of artificial intelligence, industrial Internet of Things, big data and other technologies with modern mine development. As a result, various intelligent subsystems for specific aspects of mining have emerged and been applied.

[0003] In existing technologies, most of the aforementioned intelligent construction suffers from data interoperability issues. Different systems are often independently developed by different manufacturers, employing heterogeneous data standards, communication protocols, and software platforms. This leads to several problems: First, data interoperability between systems is difficult. Safety monitoring data, equipment operation data, production scheduling data, and personnel location data cannot be effectively integrated, making it difficult to support unified perception and collaborative analysis of the overall mine status. Second, existing systems focus on monitoring and alarm functions, lacking intelligent decision-making and autonomous collaborative control capabilities based on multi-source data fusion. For example, an equipment fault diagnosis system might warn of a potential bearing malfunction in a hoist, but this information cannot automatically trigger the production scheduling system to adjust the hoisting plan, nor can it link with the video system to retrieve real-time images of the fault point for remote diagnosis. Furthermore, it is difficult to automatically generate and dispatch maintenance work orders. The entire handling process still heavily relies on manual judgment and coordination, resulting in a delayed response. Third, the lack of unified control strategies and linkage mechanisms between subsystems makes it difficult to achieve rapid, accurate, and collaborative responses across systems when dealing with complex working conditions or sudden emergencies. For example, automatic control of ventilation systems and dynamic planning and dissemination of optimal evacuation routes for personnel under disaster conditions are not feasible.

[0004] Therefore, in response to the problems mentioned above, this invention proposes an unmanned intelligent control system for mines. Summary of the Invention

[0005] To overcome the problems of data incompatibility between existing intelligent mining systems, insufficient depth of intelligent decision-making, difficulties in cross-system collaborative control, and low efficiency of emergency response, this invention proposes an unmanned intelligent management and control system for mines. This system is designed to build an integrated platform that features deep data fusion, an intelligent decision-making engine, and the ability to achieve automatic cross-business collaborative control, thereby enabling visibility, manageability, controllability, and autonomous optimization of the entire mining production process.

[0006] The technical solution of this invention is: an unmanned intelligent control system for mines, comprising: The data acquisition layer consists of various monitoring nodes deployed underground and on the surface of the mine, including vibration sensors, temperature sensors, gas sensors, video cameras, laser scanners, UWB positioning base stations, RFID readers, smart mining lamps, and wearable vital sign monitoring devices, used to collect equipment operating status, environmental parameters, personnel location, vehicle trajectory, video images, and personnel physiological data. The edge computing layer includes edge controllers deployed at key nodes in each mining area, substation and transport roadway. Each edge controller connects to several data acquisition layer devices to perform local preprocessing, feature extraction, preliminary anomaly judgment and protocol encapsulation on the acquired raw data, and communicates with the upper-layer system through industrial Ethernet ring network or 5G wireless network. The cloud-based intelligent management and control platform, built on an industrial internet architecture, includes: The data platform is used to access and integrate multi-source data from the edge computing layer and various independent subsystems, perform data governance, storage and modeling, and form a unified data asset for the mine. The independent subsystems include at least ventilation, drainage, compressed air, hoisting, transportation, power supply and safety monitoring subsystems. The intelligent analysis engine integrates algorithm models based on big data and artificial intelligence, including a BP neural network model for equipment failure prediction, a decision tree model for safety risk assessment, an optimization algorithm model for production scheduling, and a computer vision model for video analysis. The 3D visualization engine constructs a holographic digital twin scene of the mine based on mine geological data, tunnel models, and equipment models, enabling real-time dynamic mapping and visualization of personnel, equipment, and environmental data; The collaborative management and control application center includes a production scheduling collaboration module, a security assurance collaboration module, and an equipment operation and maintenance collaboration module. It is used to generate production instructions, risk warnings, and maintenance work orders based on the output results of the intelligent analysis engine, and then distribute them to the corresponding execution units or personnel terminals. The execution and control layer includes remotely controllable intelligent devices, inspection robots, emergency broadcast terminals and information display screens. It receives instructions from the cloud-based intelligent management and control platform and executes corresponding operations. The intelligent devices include variable frequency fans, automatic dampers, regulating windows, drainage pumps, hoists, belt conveyors, rock drilling rigs and loaders. The human-machine interface terminal includes a large-screen display system in the ground dispatch center, an engineer's workstation, a mobile APP for management personnel, and handheld smart terminals and wearable devices for underground personnel, which are used for system status monitoring, command issuance, alarm reception, and information query.

[0007] It is worth noting that the data acquisition layer preferably includes vibration and temperature sensors located at key bearings and housings of large fixed equipment such as main fans, air compressors, hoisting systems, and drainage pumps; ultrasonic wind speed and differential pressure sensors located in the main intake and return air roadways; and intelligent cameras and UWB positioning base stations located in areas such as mining faces, belt conveyors, and substations. The edge controller preferably supports the MQTT IoT encryption protocol and has data caching and breakpoint resume functions to ensure the integrity and continuity of data when the network is unstable.

[0008] Preferably, the intelligent analysis engine uses a BP neural network model for equipment fault prediction. This model obtains equipment vibration and temperature time-series data through a data platform, extracts features, inputs them into the BP neural network for training and health status assessment, and combines a decision tree algorithm to correct the output results, generating an equipment health assessment report and predictive maintenance recommendations.

[0009] Preferably, vibration feature extraction includes spectral analysis, envelope analysis, and cepstral analysis of the original waveform to comprehensively reflect the mechanical state of the equipment.

[0010] Preferably, the 3D visualization engine is specifically used to: integrate the 3D model of the ore body, the tunnel engineering model, the surface terrain model, and the 3D model of the main equipment to form a unified spatial digital baseboard for the mine; fuse and bind real-time collected personnel UWB positioning data, vehicle trajectory data, equipment operating status data, and environmental monitoring data to the corresponding spatial coordinates for dynamic display in the 3D scene; support query, sectioning, measurement, simulation, and analysis functions based on spatial location, and locate and highlight abnormal areas or equipment when an alarm occurs.

[0011] It is worth noting that the digital twin scenario of the mine can automatically locate and highlight abnormal areas based on real-time monitoring data (such as equipment alarms and excessive gas concentrations), providing intuitive spatial situational awareness for emergency command.

[0012] Preferably, the system also includes an intelligent ventilation control subsystem, which specifically includes wind speed and resistance monitoring sensors deployed in the roadway, remotely controllable ventilation facilities (such as automatic air doors, adjustable air windows, and frequency converters for main ventilation fans), and a ventilation network calculation module integrated into the intelligent analysis engine; wherein the collaborative management and control application center can automatically generate and issue air volume adjustment commands based on the calculation results and air demand.

[0013] Preferably, the ventilation network calculation module can dynamically calculate the ventilation network of the entire mine based on the real-time monitored ventilation parameters, and generate a linkage control scheme including main fan speed adjustment and wind window and damper opening adjustment through optimization algorithms in combination with changes in the demand of air use locations.

[0014] As a preferred embodiment, the personnel safety protection unit in the data acquisition layer specifically includes: a smart mine lamp integrating a UWB positioning module, a gas detection module, a voice communication module and a camera; a wearable smart bracelet with heart rate and blood oxygen monitoring functions, which is connected to the smart mine lamp via Bluetooth; and a physical and mental health monitoring device installed at the wellhead, which is used to automatically screen personnel for abnormal physical signs before they go down the mine. When the system detects that personnel have entered a dangerous area, exhibit abnormal vital signs, or exceed the limits for ambient gas, it will issue an audible and visual alarm through the smart mining lamp and trigger emergency broadcasts and evacuation route guidance through the collaborative management and control application center.

[0015] Preferably, the system also includes an intelligent monitoring subsystem based on AI video analysis, which specifically includes: intelligent cameras deployed along the conveyor belt, tunnel face, chute, substation and main roadway, with built-in AI processing chips; and a computer vision model integrated in the intelligent analysis engine for real-time analysis of video streams to identify conveyor belt misalignment, coal accumulation, personnel crossing illegally, not wearing safety helmets, abnormal equipment status, flames and smoke. When abnormal behavior or status is detected, an alarm event is generated, triggering on-site audible and visual alarms, and pushing alarm information, captured images, and video clips to the collaborative management application center and relevant personnel terminals.

[0016] Preferably, the system also includes an intelligent collaborative management module for ore flow, which specifically includes: real-time collection of ore output, transfer volume, and storage volume through vehicle-mounted weighing sensors, ore pass level gauges, and belt scales; and an ore flow optimization model integrated in the intelligent analysis engine, which dynamically calculates and recommends the optimal sequence of loading, transportation, crushing, and hoisting operations based on production plans, equipment status, and storage conditions. The production scheduling and coordination module of the collaborative management and control application center issues scheduling instructions or suggestions to the loader, mining truck, crusher and main conveyor equipment based on the optimization results, and dynamically displays the ore flow status in the three-dimensional visualization engine.

[0017] Preferably, the execution and control layer includes a track-mounted intelligent inspection robot, which is specifically used for: automatically inspecting the conveyor belt along a preset track, equipped with a visible light camera, an infrared thermal imager, a noise sensor, and a vibration sensor; performing local image and data analysis through an edge computing layer or its own processor to identify early signs of equipment failure; transmitting inspection data and diagnostic results back to the cloud-based intelligent management and control platform in real time, and receiving instructions from the platform for detailed inspection or emergency response.

[0018] Preferably, the edge controller of the edge computing layer has the ability to run lightweight analysis algorithms, which can perform real-time fast Fourier transform (FFT) and eigenvalue calculation on the vibration data of key equipment (large fixed equipment) to achieve millisecond-level local anomaly judgment and early warning.

[0019] Preferably, the collaborative management and control application center also includes a dual pre-control module for safety risk classification and control and accident hazard investigation and management. It automatically associates various early warning and alarm information output by the intelligent analysis engine with preset risk points and hazard databases, generates dynamic risk maps and management work orders, and automatically dispatches and tracks them through the workflow engine to achieve continuous optimization of the safety management system.

[0020] The beneficial effects of this invention are: 1. This invention solves the problem of data incompatibility between various subsystems in traditional mines and the inability to effectively integrate safety monitoring data, equipment operation data, production scheduling data and personnel location data by constructing a system architecture that integrates cloud, edge and terminal collaboration with sensing, knowledge and control. It is the first to achieve unified access, deep integration and governance of all elements of mine data.

[0021] 2. By deploying edge controllers with local computing capabilities and intelligent analysis engines in tandem, this invention achieves in-depth insight into the operating status of large fixed equipment and early prediction of faults. It transforms the maintenance mode from reactive maintenance and periodic inspections to state-based predictive maintenance, significantly improving equipment reliability, reducing unplanned downtime, and greatly enhancing economic and safety benefits.

[0022] 3. This invention realizes intelligent linkage emergency response of multiple systems. Under disaster conditions such as gas over-limit and fire, the system can automatically and in parallel perform a series of actions such as disaster location, ventilation control, personnel alarm and disaster avoidance guidance and emergency reconnaissance, reducing the response time of traditional manual serial handling from tens of minutes to seconds, and greatly improving the inherent safety level and disaster resistance of mines. Attached Figure Description

[0023] Figure 1 The diagram shown is a schematic representation of the system framework of this invention. Figure 2 The diagram illustrates the predictive maintenance and production collaboration process for large fixed equipment according to the present invention. Figure 3 The diagram shown is a schematic of the intelligent ventilation and disaster emergency linkage control process of the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but 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.

[0025] Please see Figure 1 This invention provides an embodiment: an unmanned intelligent control system for mines. In this embodiment, the data acquisition layer fully covers key areas such as the underground mining face, main transport roadways, electromechanical chambers, and surface industrial areas. It is scientifically deployed according to the "Data Resource Planning" table in the "Non-Coal Mine Digital Mine Construction Plan" and the "Equipment Measurement Point Layout Table" in the "Research on Key Technologies for Intelligent Diagnosis of Safe Operation of Large Fixed Equipment in Coal Mines Based on Big Data". For example, at the drive and non-drive bearings of the main fan motor and the fan impeller bearings, acceleration sensors and platinum resistance temperature sensors for monitoring radial and axial vibration are installed simultaneously; ultrasonic wind speed sensors and differential pressure sensors are installed at the main intake and return air roadway wind measurement stations; integrated vibration and temperature sensors are installed at the drive and redirection roller bearings of the belt conveyor, and an intrinsically safe thermal imager is deployed on the non-load side of the rollers to monitor the roller surface temperature field; all personnel and vehicles are equipped with integrated UWB positioning modules. The smart mining lamp, equipped with a gas detection module, voice communication module, and camera, is wirelessly connected to a smart wristband that monitors heart rate and blood oxygen. This layer enables real-time, synchronous, and high-precision acquisition of information from all elements, including "people (location, status), machines (vibration, temperature, current, video), environment (gas, wind speed, temperature, dust), and management (instructions, processes)."

[0026] In this embodiment, the edge computing layer includes edge controllers installed in key on-site areas such as mining substations, conveyor heads, and pump rooms. Their function is to achieve local data processing and intelligent data forwarding. For example, the edge controller connected to the main drainage pump group's sensors can perform real-time Fast Fourier Transform (FFT) on multiple vibration signals, calculate the effective value of velocity, peak value, and kurtosis, and compare them with a built-in threshold model. Once the characteristic value exceeds the warning line, a primary alarm is immediately generated locally, and the event is reported through a low-latency industrial ring network, while high-fidelity raw waveform data is cached. This changes the previous model where all raw data was uploaded, and the analysis burden was concentrated on a central server. The edge controller supports IoT encryption protocols such as MQTT and has data caching and breakpoint resumption mechanisms to ensure that critical data is not lost during network interruptions or congestion, and automatically resumes transmission after network recovery, ensuring the robustness of the data link.

[0027] In this embodiment, the cloud-based intelligent management and control platform is built based on the concept of an industrial internet middleware platform. The data middleware platform first accesses, cleans, standardizes, and integrates massive amounts of data from the edge layer and various independent subsystems. It correlates real-time vibration data of equipment, current and voltage data of the power grid, methane concentration data of the environment, and personnel location trajectory data under a unified spatiotemporal reference (based on the mine's three-dimensional geographic coordinate system and nanosecond-level time synchronization) to form a database describing the overall operating status of the mine. The intelligent analysis engine is the core intelligence of the platform, with built-in containers of various configurable and trainable algorithm models. For example, for predictive maintenance of large fixed equipment, a time-series prediction model based on the fusion of deep belief networks and long short-term memory networks is installed. Its input is the multi-dimensional time-series characteristics of equipment (vibration spectrum characteristics, temperature trends, and associated motor current harmonics) spanning multiple days, which have been standardized by the data middleware platform. The output is the probability distribution of the remaining useful life (RUL) of key components (such as bearings and gears) and the confidence level of specific failure modes in the next 24-72 hours. The 3D visualization engine, based on data fusion from the data platform, drives a high-precision digital twin model of the mine. This model not only statically displays the 3D spatial relationships of roadways, equipment, and pipelines, but also dynamically maps them. It uses flashing red dots to indicate the location of alarm equipment, different colored streamlines to dynamically display the airflow direction and speed in the roadways, and moving light spots to display the real-time trajectories of personnel and vehicles. The collaborative management and control application center carries the specific business logic. It receives the conclusions from the intelligent analysis engine, combines them with production plans, safety procedures, and resource status, and automatically generates and dispatches collaborative task chains. For example, it automatically generates maintenance work orders that include faulty equipment, suggested maintenance content, required spare parts, estimated working hours, and safety measures, and pushes them to the mobile terminals of electromechanical engineers. At the same time, it sends a reminder to the production scheduling module to "adjust and improve the plan to reserve a time window for maintenance."

[0028] In this embodiment, the execution and control layer receives and executes precise instructions from the collaborative management and control application center. This includes two main categories: first, direct digital command control of existing remotely controllable equipment (such as main ventilation fans equipped with frequency converters, electrically adjustable windows, and PLC-controlled drainage pump sets); second, expanding the control boundary by deploying autonomous equipment such as intelligent inspection robots and drones. For example, after receiving the instruction to "conduct a detailed inspection of the drive unit of conveyor belt No. 3," the track-mounted intelligent inspection robot automatically plans its path to the designated location, adjusts the angle of the infrared thermal imager using its robotic arm, performs multi-angle temperature measurements on the bearing housing, and collects high-definition video. All data is transmitted back in real time and sent to the information publishing terminal.

[0029] In this embodiment, the human-machine interface terminal provides a unified interface for users with different roles. The ground dispatch center displays a 3D overall mine situation, key monitoring videos, and production indicator charts on a large screen. Management personnel receive alarm push notifications, access approval processes, and view reports via a mobile app. On-site inspection personnel receive work orders, scan equipment QR codes to obtain historical files, and upload inspection results via an engineer's workstation. All terminals access the unified platform through single sign-on, and data views and operation permissions are dynamically configured according to roles.

[0030] This invention provides Embodiment 1: Please see Figure 2 This embodiment describes the initial inspection and maintenance process for large fixed equipment. Specifically: (1) A vibration sensor installed on the high-speed shaft bearing of the gearbox of the auxiliary shaft hoist continuously collects vibration signals at a sampling rate of 12.8kHz. The local edge controller calculates the effective value and peak value of the vibration velocity every 5 seconds and monitors continuously. If the edge controller finds that the effective value of the vibration velocity increases significantly within 1 hour, it immediately generates a primary event alarm of "increased vibration of the gearbox of the auxiliary shaft hoist" and reports it to the cloud platform through the industrial ring network, along with the raw waveform data packets of the most recent 10 minutes.

[0031] (2) The data platform correlates the received waveform data with historical data of the same hoist and normal operation data of similar hoists. The equipment health management model in the intelligent analysis engine is triggered to perform fine spectrum analysis, demodulation analysis and cepstral analysis on the original waveform. The analysis found that there was a significant and continuously increasing sideband at the gear meshing frequency, which was highly consistent with the characteristics of early tooth breakage failure. The model combined temperature trend and load history to calculate that the probability of tooth breakage failure of the gear in the next 7 days is 85%, and outputs a diagnostic report: "Suspected local pitting corrosion of gear, it is recommended to arrange an unpacking inspection as soon as possible".

[0032] (3) The production scheduling module and equipment operation and maintenance module of the collaborative management and control application center receive the diagnostic report synchronously. The operation and maintenance module automatically retrieves equipment files and spare parts inventory, generates a "preventive maintenance work order", designates the person in charge, suggests maintenance time and requires special operation permits. The production scheduling module evaluates the current lifting task and the status of the standby lifting system, automatically calculates a 4-hour maintenance time window with the least impact on output, and associates this window suggestion with the maintenance work order.

[0033] (4) The associated "collaborative task package" is pushed to the terminals of the electromechanical section chief and the dispatch director for final confirmation. After approval, the system automatically dispatches specific work orders and safety measures to the smart terminals of the maintenance team members and sends a planned shutdown notice to the hoist operator's terminal. After the maintenance is completed, the maintenance personnel upload maintenance photos and spare parts replacement information through the terminal. The system automatically updates the equipment file and records the entire process data from early warning to closed loop, which is used to optimize the early warning threshold of the future model.

[0034] This embodiment demonstrates that transforming reactive or scheduled maintenance into precise predictive maintenance, through in-depth intelligent analysis of vibration signals, can improve fault identification accuracy from approximately 70% using traditional manual spectrum analysis to over 90%. Through automated system collaboration, the time from anomaly detection to maintenance plan generation is reduced from several hours in the traditional method to less than 30 minutes, and unplanned downtime is expected to be reduced by 40%-60%.

[0035] This invention provides Embodiment 2: Please see Figure 3 This embodiment describes the control process for a single intelligent ventilation system: (1) The system uses sensors deployed at the mining face and return airway to monitor the air volume, wind speed, etc. in real time. , Parameters such as temperature are collected, and the data platform integrates these data with the 3D model of the ventilation system (including roadway wind resistance characteristics and ventilation facility status). The real-time ventilation network calculation module in the intelligent analysis engine runs on a 1-minute cycle to dynamically calculate the ventilation network of the entire mine and obtain the current actual air volume distribution of each branch.

[0036] (2) When the system detects that the CO concentration in the return air corner of a certain intelligent coal mining face increases from 15ppm to 22ppm (close to the 24ppm alarm value) due to the accelerated coal cutting by the coal mining machine, the collaborative control application center determines, based on the calculation model and preset rules, that the air volume of the working face needs to be increased by about 15%. Then the system generates a sequence of control instructions: First, fine-tune the regulating window of the return air roadway of the working face, increasing the opening by 10%; Second, if the air volume is still insufficient, fine-tune the upstream associated air doors in sequence according to the preset logic; Third, simultaneously calculate the impact on the air volume distribution and energy consumption of the entire mine if the frequency of the main ventilation fan is increased from 45Hz to 47Hz. After confirmation by the dispatcher, the instructions are issued through the control layer and executed automatically.

[0037] As can be seen from this embodiment, the present invention realizes the change from constant air volume supply to dynamic air supply on demand. Under the premise of ensuring safety, the average operating frequency of the main ventilation fan can be reduced by 5-10%.

[0038] This invention provides embodiment 3: This example illustrates the process for controlling unsafe acts by personnel: (1) When a worker wearing a smart miner's lamp enters a UWB electronic fence marked as a "temporary blasting warning zone", the system automatically identifies this "personnel intrusion into a high-risk area" behavior. At the same time, the AI ​​video camera in the area identifies another person who is not wearing a safety helmet. The intelligent analysis engine links these two pieces of information in real time with whether there is a "blasting operation" plan in the area. Based on this, the dual pre-control module of the collaborative management and control application center automatically generates a "comprehensive hidden danger": "An unauthorized person has intruded into the blasting warning zone of the tunnel without wearing a safety helmet, which is associated with the risk of blasting operation, level: major".

[0039] (2) The hazard is automatically dispatched to the mobile terminals of the on-site safety officer, the blasting operation supervisor, and the team management personnel via the workflow engine. Upon arrival at the site, the safety officer confirms the evacuation of personnel and the implementation of safety measures by taking photos via their terminal, and uploads the results. The system automatically records the entire process and time taken. If the hazard is not rectified within the specified time limit, the system automatically escalates and pushes it to a higher-level manager. All violations and hazard data are automatically linked to the responsible personnel and work teams, serving as a quantitative basis for assessing the safety production responsibility system.

[0040] When the smart bracelet of an underground worker detects that a worker's heart rate consistently exceeds 120 beats per minute and blood oxygen saturation is below 90%, the system's health monitoring module issues an "abnormal physiological state" warning and pushes it to the shift leader and the surface health administrator. The administrator can then immediately conduct a video call with the worker's smart miner's lamp through the system to inquire about the situation and, if necessary, instruct the worker to evacuate from the work area.

[0041] As demonstrated in this embodiment, the present invention links discrete violations, hidden dangers, and alarm information through spatiotemporal and business logic to form a systematic risk profile, thereby reducing the rate of missed hazard reports. It achieves a closed-loop, automated, and traceable safety management process, and the average rectification time for hidden dangers is expected to be shortened by 30%. Personnel vital sign monitoring provides technical support for preventing overwork and sudden illnesses.

[0042] Comparative Example 1 provided by the present invention: This example simulates early fatigue damage to the rolling bearing of a main drainage pump. Verification is conducted through comparative experiments over a 6-month evaluation period. The comparative example uses an independent vibration monitoring system installed on the water pump motor. This system only monitors the total vibration value, and the alarm threshold is set to a fixed value. When the vibration slowly increases to the threshold and triggers an alarm, maintenance personnel need to manually analyze the historical spectrum to determine the type and severity of the fault, and then coordinate the downtime.

[0043] The embodiment of this invention involves a system that continuously analyzes vibration spectrum characteristics through edge computing. Even before the total vibration value reaches a traditional threshold, the system identifies early fault characteristics using an intelligent model, issues predictive warnings in advance, and automatically links maintenance resources with production plans.

[0044] Table 1 Comparison of Timeliness and Accuracy of Fault Early Warning for Large Fixed Equipment

[0045] As shown in the table above, firstly, the comparative model can only trigger alarms based on fixed thresholds 2 to 24 hours before a fault occurs, while the present invention, through real-time spectrum feature analysis on the edge side and in-depth mining of cloud-based intelligent models, successfully achieves early predictive warnings on average 3 to 15 days in advance, increasing the warning lead time by an order of magnitude; secondly, the comparative model relies on manual interpretation of the spectrum, with a diagnostic accuracy rate hovering around 65%, limited by personal experience, while the present invention uses an artificial intelligence model that fuses multi-dimensional features, steadily improving the accuracy rate of automatic fault type diagnosis to over 90%, significantly reducing misjudgments and missed judgments; thirdly, in the traditional model, it takes an average of 4 to 8 hours of manual communication and coordination from alarm to maintenance decision-making, while the present invention's system, through built-in collaborative management logic, automates and compresses this process to within 1 hour, ultimately reducing unplanned downtime events by about 55%, and increasing the inventory turnover rate of maintenance spare parts by about 25% through accurate demand forecasting.

[0046] Comparative Example 2 is provided in this invention: This example simulates a sudden exceedance of the gas sensor limit in the return airway of a coal mining face. Verification is performed through comparative experiments, including: In the comparative scenario, the safety monitoring system alarms, and the dispatcher notifies the ventilation and production departments by phone. The ventilation technician manually calculates the required increase in airflow based on experience and instructs the underground gas inspector to adjust the ventilation windows via telephone. Personnel positioning and video monitoring are independent systems and must be switched for viewing.

[0047] The embodiment of the present invention is that the system automatically locates the over-limit point, retrieves surrounding video, and automatically generates and sends air-increasing control commands to the corresponding automatic windows based on real-time ventilation network calculations. At the same time, based on personnel location information, it automatically notifies and guides personnel in the over-limit area to evacuate to the upwind side.

[0048] Table 2 Comparison of Emergency Response and Ventilation Control Efficiency for Gas Exceedance Limits

[0049] As shown in the table above, this invention reduces the response time of 8-15 minutes required for manual reporting, communication, judgment, and command in the traditional mode to an automated response of 30-60 seconds, increasing the speed by more than 10 times. At the same time, through precise calculation and automatic control based on a real-time ventilation network model, the time for the gas concentration to decrease from the excessive limit to the safe value is reduced from the traditional 15-30 minutes to 5-10 minutes, improving the handling efficiency by 2-3 times. This invention realizes comprehensive and rapid alarm and path guidance for personnel in the affected area through broadcasting and smart mine lamps, and automatically links and archives the instructions, data, videos, and location information of the entire event process, providing a complete and reliable digital record for post-event analysis and traceability.

[0050] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A mine unmanned intelligent control system, characterized in that, Including: The data acquisition layer consists of various monitoring nodes deployed underground and on the surface of the mine, including vibration sensors, temperature sensors, gas sensors, video cameras, laser scanners, UWB positioning base stations, RFID readers, smart mining lamps, and wearable vital sign monitoring devices, used to collect equipment operating status, environmental parameters, personnel location, vehicle trajectory, video images, and personnel physiological data. The edge computing layer includes edge controllers deployed at key nodes in each mining area, substation and transport roadway. Each edge controller connects to several data acquisition layer devices to perform local preprocessing, feature extraction, preliminary anomaly judgment and protocol encapsulation on the acquired raw data, and communicates with the upper-layer system through industrial Ethernet ring network or 5G wireless network. The cloud-based intelligent management and control platform, built on an industrial internet architecture, includes: The data platform is used to access and integrate multi-source data from the edge computing layer and various independent subsystems, perform data governance, storage and modeling, and form a unified data asset for the mine. The independent subsystems include at least ventilation, drainage, compressed air, hoisting, transportation, power supply and safety monitoring subsystems. The intelligent analysis engine integrates algorithm models based on big data and artificial intelligence, including a BP neural network model for equipment failure prediction, a decision tree model for safety risk assessment, an optimization algorithm model for production scheduling, and a computer vision model for video analysis. The 3D visualization engine constructs a holographic digital twin scene of the mine based on mine geological data, tunnel models, and equipment models, enabling real-time dynamic mapping and visualization of personnel, equipment, and environmental data; The collaborative management and control application center includes a production scheduling collaboration module, a security assurance collaboration module, and an equipment operation and maintenance collaboration module. It is used to generate production instructions, risk warnings, and maintenance work orders based on the output results of the intelligent analysis engine, and then distribute them to the corresponding execution units or personnel terminals. The execution and control layer includes remotely controllable intelligent devices, inspection robots, emergency broadcast terminals and information display screens. It receives instructions from the cloud-based intelligent management and control platform and executes corresponding operations. The intelligent devices include variable frequency fans, automatic dampers, regulating windows, drainage pumps, hoists, belt conveyors, rock drilling rigs and loaders. The human-machine interface terminal includes a large-screen display system in the ground dispatch center, an engineer's workstation, a mobile APP for management personnel, and handheld smart terminals and wearable devices for underground personnel, which are used for system status monitoring, command issuance, alarm reception, and information query.

2. The mine unmanned intelligent management and control system according to claim 1, characterized in that, The construction and operation of the BP neural network model for equipment fault prediction in the intelligent analysis engine are as follows: A1. The data platform acquires historical and real-time vibration and temperature time-series data of large fixed equipment and their corresponding equipment health status tags. Large fixed equipment includes main fan, air compressor, main / auxiliary lifting system and main drainage pump. A2 extracts multidimensional feature values ​​by performing waveform spectrum analysis, envelope analysis, and cepstral analysis on vibration data. A3. The extracted feature values ​​are input into the equipment fault prediction BP neural network for training. The output layer nodes of the equipment fault prediction BP neural network correspond to the health status score or the probability of a specific fault mode of the equipment. A4, combined with the decision tree algorithm, corrects and interprets the output of the neural network, and finally generates equipment health assessment reports and predictive maintenance suggestions, and pushes maintenance work orders through the equipment operation and maintenance collaboration module of the collaborative management and control application center.

3. The unmanned intelligent control system for mines according to claim 1, characterized in that, The 3D visualization engine is specifically used to: integrate the 3D model of the ore body, the tunnel engineering model, the surface terrain model, and the 3D model of the main equipment to form a unified spatial digital baseboard for the mine; fuse and bind real-time collected personnel UWB positioning data, vehicle trajectory data, equipment operating status data, and environmental monitoring data to the corresponding spatial coordinates for dynamic display in the 3D scene; support query, sectioning, measurement, simulation, and analysis functions based on spatial location, and locate and highlight abnormal areas or equipment when an alarm occurs.

4. The unmanned intelligent control system for mines according to claim 1, characterized in that, The ventilation control subsystem of the system specifically includes: Ultrasonic wind speed sensors and differential pressure sensors deployed in the main intake and return airway are used to monitor the air volume and air resistance of the roadway in real time; remotely controllable automatic dampers, regulating windows and main ventilation fan frequency converters; ventilation network calculation module and on-demand air supply optimization algorithm integrated into the intelligent analysis engine; The security and coordination module of the collaborative management and control application center generates and issues air volume adjustment instructions to the corresponding automatic air doors, regulating windows and main ventilation fans based on the calculation results and the air demand of the air usage points, so as to realize the dynamic and precise control of the total air volume and local air volume of the mine.

5. The unmanned intelligent control system for mines according to claim 1, characterized in that, The personnel safety protection unit in the data acquisition layer specifically includes: a smart mine lamp integrating a UWB positioning module, a gas detection module, a voice communication module and a camera; a wearable smart bracelet with heart rate and blood oxygen monitoring functions, which is connected to the smart mine lamp via Bluetooth; and a physical and mental health monitoring device installed at the wellhead, which is used to automatically screen personnel for abnormal physical signs before they go down the mine. When the system detects that personnel have entered a dangerous area, exhibit abnormal vital signs, or exceed the limits for ambient gas, it will issue an audible and visual alarm through the smart mining lamp and trigger emergency broadcasts and evacuation route guidance through the collaborative management and control application center.

6. The unmanned intelligent control system for mines according to claim 1, characterized in that, The system also includes an intelligent monitoring subsystem based on AI video analysis, which specifically includes: intelligent cameras deployed along the conveyor belt, tunnel face, chute, substation and main roadway, with built-in AI processing chips; and a computer vision model integrated in the intelligent analysis engine, used to analyze video streams in real time and identify conveyor belt deviation, coal accumulation, personnel crossing illegally, not wearing safety helmets, abnormal equipment status, flames and smoke. When abnormal behavior or status is detected, an alarm event is generated, triggering on-site audible and visual alarms, and pushing alarm information, captured images, and video clips to the collaborative management application center and relevant personnel terminals.

7. The unmanned intelligent control system for mines according to claim 1, characterized in that, The system also includes an intelligent collaborative management module for ore flow, which specifically includes: real-time collection of ore output, transfer volume and storage volume through vehicle-mounted weighing sensors, ore pass level gauges and belt scales; and an ore flow optimization model integrated in the intelligent analysis engine, which calculates and recommends the optimal sequence of loading, transportation, crushing and hoisting operations based on production plans, equipment status and storage conditions. The production scheduling and coordination module of the collaborative management and control application center issues scheduling instructions or suggestions to the loader, mining truck, crusher and main conveyor equipment based on the optimization results, and dynamically displays the ore flow status in the three-dimensional visualization engine.

8. The mine unmanned intelligent management and control system according to claim 1, characterized in that, The execution and control layer includes a track-mounted intelligent inspection robot, which is specifically used to: automatically inspect the conveyor belt along a preset track, equipped with a visible light camera, an infrared thermal imager, a noise sensor, and a vibration sensor; perform local image and data analysis through the edge computing layer or its own processor to identify early signs of equipment failure; and transmit the inspection data and diagnostic results back to the cloud-based intelligent management and control platform in real time, and receive instructions from the platform for detailed inspection or emergency response.

9. The mine unmanned intelligent management and control system according to claim 1, characterized in that, The edge controller of the edge computing layer has the following functions: Supports IoT encryption protocols to enable secure data communication with cloud platforms and peer controllers; Built-in data caching and breakpoint resume mechanism: data is stored locally when the network is interrupted and automatically resumed when the network is restored; A lightweight analysis algorithm is used to perform real-time FFT transformation and eigenvalue calculation on vibration data of key equipment, enabling millisecond-level local anomaly detection and early warning.

10. The mine unmanned intelligent management and control system according to claim 1, characterized in that, The collaborative management and control application center also includes a dual-prevention module for safety risk classification and control and accident hazard investigation and management, which is specifically used for: B1 automatically associates the equipment fault warnings, environmental over-limit alarms, video violation identification, and abnormal personnel vital signs information output by the intelligent analysis engine with the preset risk points and hidden danger database, generating dynamic risk maps and hidden danger management work orders; B2, through the workflow engine, automatically dispatches governance work orders to the handheld terminals or mobile apps of the relevant responsible personnel, and tracks the closed-loop processing. B3, based on historical data and processing results, continuously optimizes the risk identification database and hazard judgment model to achieve the autonomous evolution of the safety management system.

Images