Fully-mechanized coal mining equipment interlocking control method based on AI identification and sensor fusion feedback

Abstract

The invention discloses a fully-mechanized coal mining equipment interlocking control method based on AI identification and sensor fusion feedback. The method specifically comprises the following steps that 1, underground coal mining and transportation video images are collected in real time; 2, a sensor is installed underground to monitor the operation condition of fully-mechanized coal mining equipment and environmental indexes in a stope space; step 3, accessing the AI intelligent identification module in the step 1 and the monitoring data in the sensor in the step 2 to a big data platform; 4, the abnormal condition of the underground operation site is recognized through an AI intelligent recognition module; step 5, according to the abnormal condition identified in the step 4 and the monitoring data of the sensor, issuing control instructions of different early warning levels to a fully-mechanized coal mining equipment control system; and step 6, performing linkage control on the fully-mechanized coal mining equipment according to the control instructions of different grades issued in the step 5, and realizing an interlocking control protection mechanism of the fully-mechanized coal mining equipment. The problems that in the prior art, a fully mechanized coal mining face is single in early warning mode, and potential safety hazards exist are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of coal mine intelligence, and relates to an interlocking control method for fully-mechanized mining equipment based on AI recognition and sensor fusion feedback. Background Art

[0002] The intelligent system underground in coal mines is formed by numerous subsystems, large and small, operating simultaneously, such as communication, sensing, and regulation. In order to effectively coordinate the fully-mechanized mining automation system, achieve automatic and intelligent mining of the working face, reduce the possibility of safety accidents as much as possible, and ensure safe and efficient mining of the working face; currently, the early warning method for intelligent fully-mechanized mining faces is single, only through a single sensor for early warning (such as personnel overrun, gas, etc.). At present, there are problems in the intelligent construction of coal mines, such as inconsistent data models, incomplete sensor data, and cross-regional data fusion. Moreover, the predicted risk factors are relatively single, the effect of the single early warning method is poor and relatively one-sided, resulting in frequent occurrence of safety hazards such as faults of underground fully-mechanized mining equipment and increased hidden risk factors during the intelligent mining process of the working face. Summary of the Invention

[0003] The purpose of the present invention is to provide an interlocking control method for fully-mechanized mining equipment based on AI recognition and sensor fusion feedback, which solves the problems of single early warning method and large safety hazards existing in the prior art for fully-mechanized mining faces.

[0004] The technical solution adopted by the present invention is an interlocking control method for fully-mechanized mining equipment based on AI recognition and sensor fusion feedback, which specifically includes the following steps: Step 1, collect real-time video images of underground coal mining and transportation, and store and transmit the obtained video stream into the AI intelligent recognition module; Step 2, monitor the operating conditions of fully-mechanized mining equipment and environmental indicators in the stope space by installing sensors underground; Step 3, connect the monitoring data in the AI intelligent recognition module in Step 1 and the sensors in Step 2 to the big data platform; Step 4, identify abnormal situations at the underground operation site through the AI intelligent recognition module; Step 5, issue control instructions with different warning levels to the control system of fully-mechanized mining equipment based on the abnormal situations identified in Step 4 and the monitoring data of the sensors; Step 6, perform linkage control on the fully-mechanized mining equipment according to the different-level control instructions issued in Step 5 to implement the interlocking control protection mechanism for fully-mechanized mining equipment.

[0005] The characteristics of the present invention also lie in: In Step 1, video images of underground coal mining and coal transportation are obtained by installing cameras underground.

[0006] The sensors in Step 2 include carbon monoxide sensors, methane sensors, and dust sensors.

[0007] The warning levels in Step 5 are as follows: Level I: Pop-up warning on the PC side of the big data intelligent mining platform, no handling; Level II: Pop-up warning - audible and visual alarm of the underground lighting alarm device - voice broadcast prompt in the ground dispatching room and underground; Level III: Pop-up warning - audible and visual alarm - voice broadcast prompt - one-key start and stop of fully-mechanized mining equipment.

[0008] The specific process of Step 5 is as follows: Step 5.1, when the AI intelligent recognition module recognizes that there is foreign matter on the belt and it does not affect the normal operation of the belt, a pop-up warning is given, no handling is done, and a Level I control instruction is issued; when the AI intelligent recognition module recognizes that there is foreign matter on the belt and it affects the normal operation of the belt or even causes damage to the belt, the equipment control system can automatically give an audible and visual alarm, voice prompt, issue a Level III control instruction, and implement one-key start and stop of the shearer; Step 5.2, if the AI intelligent recognition module recognizes that the operating parameters of the shearer, the operating parameters of the support, the operating parameters of the crusher - conveyor - scraper conveyor, and the operating parameters of the pump station are abnormal and affect the normal operation of the fully-mechanized mining equipment, a Level III warning is issued, and the system gives a pop-up warning - audible and visual alarm - voice broadcast prompt - one-key start and stop of the shearer; Step 5.3, when the AI intelligent recognition module recognizes that the hydraulic support has problems such as untimely support moving and untimely retraction of the rib protection plate, which affect the normal mining of the working face, a Level III warning is given, and one-key start and stop of the shearer is implemented; Step 5.4, when the AI intelligent recognition module recognizes that the scraper chain of the scraper conveyor has cracks and affects the normal operation of the scraper conveyor, a device danger warning is given, one-key start and stop operation of the shearer is carried out, and a Level III control instruction is issued; Step 5.5, using the real-time video data of the shearer following the shearer operation in the fully-mechanized mining face captured by the hydraulic support camera, when the AI intelligent recognition module recognizes that there are people entering the dangerous area of the fully-mechanized mining face, a pop-up warning, audible and visual alarm, and voice prompt are given until the people leave this area, and a Level II control instruction is issued; Step 5.6, when the sensors detect that the methane concentration exceeds 1.00%, the carbon monoxide concentration exceeds 24 ppm, and the dust concentration exceeds 1.5 mg / m 3 3, immediately a pop-up warning on the platform is given, audible and visual alarm and voice prompt underground, one-key start and stop of underground fully-mechanized mining equipment, the dispatching room commands underground personnel to evacuate in an orderly manner, and a Level III control instruction is issued; Step 5.7, when the number of personnel in the working face reaches the upper limit, a voice prompt is given. If there are subsequent personnel entering the working face area, the big data platform gives a pop-up warning, issues a Level II warning, gives an audible and visual warning underground, and voice alarm until the extra personnel in the working face evacuate from this area.

[0009] In step 5.2, abnormal operation parameters of the shearer include: Abnormal current: measured current value > rated current value; Abnormal voltage: measured voltage value > rated voltage value or measured voltage value < allowable minimum voltage value; Abnormal load: load rate > rated load rate.

[0010] In step 5.2, abnormal operation parameters of the support include: Abnormal hydraulic pressure: measured hydraulic pressure value > maximum allowable pressure value or measured hydraulic pressure value < minimum allowable pressure value; Abnormal support moving speed: measured moving speed > maximum allowable speed or measured moving speed < minimum allowable speed.

[0011] In step 5.2, abnormal operation parameters of the crusher - conveyor - scraper conveyor include: Abnormal vibration: vibration amplitude > allowable maximum vibration amplitude; Abnormal rotation speed: measured rotation speed < allowable minimum rotation speed or measured rotation speed > allowable maximum rotation speed.

[0012] The beneficial effects of the present invention are as follows: (1) AI fusion recognition and hierarchical linkage control: Through the fusion of AI intelligent recognition technology and multi - sensor data feedback, the present invention divides effective hierarchical classification linkage control strategies. This strategy not only improves the system's adaptability to complex environments but also ensures the precise execution of control instructions, providing a solid guarantee for safe production.

[0013] (2) Diversified early warning and intelligent interlocked mining: The present invention introduces an intelligent interlocked mining system, realizing the diversification of early warning indicators for the fully - mechanized mining face. Compared with the previous single early warning indicator, this system can capture potential risks in the production process more comprehensively and accurately, so as to take preventive measures in time to avoid accidents.

[0014] (3) Real - time dynamic monitoring and efficient emergency response: The present invention has the ability to perform real - time dynamic monitoring of the operating conditions of on - site equipment and the safety behaviors of personnel, and can quickly form effective combined control strategies. Through intelligent control collaborative services, the system can implement multiple protection measures such as audible and visual warnings, voice broadcasts, and interlock protection for equipment, ensuring that in the face of complex production safety emergency requirements, it can respond efficiently, quickly, and flexibly to ensure production safety. Description of the Drawings

[0015] Figure 1 is the AI recognition flowchart in the fully - mechanized mining equipment interlock control method based on AI recognition and sensor fusion feedback of the present invention; Figure 2 Flowchart of foreign object recognition in the integrated mining equipment interlock control method based on AI recognition and sensor fusion feedback of the present invention. Specific implementation manners

[0016] The present invention will be described in detail below with reference to the accompanying drawings and specific implementation manners.

[0017] Example 1 The intelligent mining interlock control platform adopted by the integrated mining equipment interlock control method based on AI recognition and sensor fusion feedback of the present invention can be connected to the main equipment control systems of underground mining and transportation recognition scenarios. According to the safety control strategies formed by the intelligent decision-making analysis management platform, it converts the combined control classification strategies recognizable by the equipment and issues different levels of linkage control instructions to the equipment control systems to achieve linkage control according to scene anomalies.

[0018] Example 2 The integrated mining equipment operation interlock protection control method (intelligent mining interlock control system) adopted by the integrated mining equipment interlock control method based on AI recognition and sensor fusion feedback of the present invention mainly includes two major parts, namely the AI intelligent recognition module and the interlock control module. The specific functions of each part are as follows: 1) AI intelligent recognition module The AI intelligent recognition module mainly includes monitoring the actual situation underground (video monitoring, sensor monitoring), and performing real-time annotation of the positions of various equipment and personnel in the video monitored by video monitoring. When there are situations of unsafe behaviors of people and unsafe states of objects, the AI intelligent recognition module will publish the specific time and risk level identified to the intelligent interlock control module.

[0019] The video stream is accessed through a dedicated network to the video server for transfer. At the same time, multiple GPU servers asynchronously execute feature extraction and trajectory calculation of equipment and personnel in the form of task slots according to different scenario videos. By combining the final recognition identifier with AR enhanced visualization technology, the three-way fusion of the AI recognition identifier, equipment working condition data, and video stream is realized, and it is transmitted to the video server for storage. The virtual reality is restored through AI recognition and AR enhanced visualization on the system video interface.

[0020] 2) Intelligent interlock control module According to the specific events and risk levels fed back by the AI intelligent recognition module, as well as the equipment data monitored by sensors, it judges the equipment operating conditions and the execution degree of the production process, and issues deceleration and stop instructions to the equipment in the dangerous area to carry out interlock protection and achieve intelligent linkage between various production systems.

[0021] Relying on the camera systems arranged in multiple scenarios and the mine's own environmental monitoring, transportation, ventilation, personnel positioning, network, fully mechanized mining and other systems, the equipment, personnel and environment in each area are monitored in real time, and information such as underground personnel, equipment, environment and ground security is fed back in real time, achieving all-round regional monitoring and control of security layout, safety protection, equipment status, environment, etc.

[0022] The intelligent mining interlocking control platform is based on the big data decision-making center of the IMS platform, integrating the collection of industrial equipment in multiple mine scenarios, various protocols, and the drivers of each device, and realizing the parsing of multiple device protocols. (For example, docking with the personnel positioning system to achieve the positioning of personnel violations; docking with the environmental detection system to realize the function of the linkage system broadcasting an alarm when the methane and dust concentrations exceed the warning values; docking with the face broadcast system and the intelligent system to realize the linkage of core production equipment; and docking with the main haulage and auxiliary haulage systems to realize system linkage, ensuring the safe production operation of the mine in multiple aspects and scenarios.) Integrate the data access of each device and system, integrate the data of each API interface, file system, and streaming data access, and perform escape according to the standard format. Support data access under various industrial protocols such as Modbus, CAN, EtherCAT, EthernetIP, OPC UA, etc., complete data parsing, verification, conversion, and the two-way data interaction of sending the control data after escape.

[0023] Embodiment 3 The fully mechanized mining equipment interlocking early warning protection control method based on AI intelligent recognition and sensor fusion feedback of the present invention has a process as Figure 1 shown, and is specifically realized through the following steps: S1. Through the cameras installed at each key position in the underground coal mining and coal transportation sites, the actual situations of underground coal mining and transportation are monitored in real time, and the obtained video streams are stored and transmitted to the AI intelligent recognition module; S2. Through various types of sensors installed underground (monitoring carbon monoxide concentration through carbon monoxide sensors, monitoring methane concentration through methane sensors, monitoring dust concentration through dust sensors, monitoring air volume through air volume sensors, and monitoring whether the number of personnel in the working face exceeds the limit through personnel positioning sensors), the operating conditions of fully mechanized mining equipment and environmental indicators in the stope space are monitored; S3. Through the intelligent big data platform, the video monitoring system (AI recognition module) and sensor monitoring data are connected to the big data platform; S4. Through the cameras installed underground, the AI intelligent recognition module identifies abnormal situations at the coal mining and transportation production operation sites according to methods such as image recognition and image comparison. For example, if it is identified that underground workers are absent from their posts during coal mining work, the AI video will monitor and intercept this unsafe behavior.

[0024] S5. Through intelligent mining interlock control, different levels (I, II, III) of control instructions are sent to the equipment control system based on the equipment operation conditions, identification of personnel's unsafe behaviors, and sensor monitoring data.

[0025] S6. Through the above - divided safety and danger levels, when the AI intelligent recognition system and sensors identify and monitor the unsafe behavior of personnel being absent from their posts, a level - II control instruction is issued. Through the intelligent interlock control system, methods such as system pop - up prompts and voice announcements can be used to remind the absent personnel to return to their posts.

[0026] Example 4 The monitoring of the machine operation status needs to complete the following steps: collect the operating parameters related to the machine operation status, perform corresponding processing on the collected data to obtain the characteristic parameters that play a decisive role in the machine operation status, judge and identify the parameters that exceed the limit state or deviate from the normal standard change, analyze and determine the cause and nature of the fault, and predict the time of component failure and maintenance measures.

[0027] Among them, the equipment information of the fully - mechanized mining face includes the operation status information of the shearer, the operation parameters of the support, the operation parameters of the crusher - conveyor - coal winning conveyor, and the operation parameters of the pump station.

[0028] a. The operation parameters of the shearer include the operating current and temperature of the working motor, the temperature of the rocker arm shaft, the drum height and undercut amount, the traveling speed of the shearer, the positioning position of the shearer, the pitching mining angle of the shearer, the dip angle of the working face in the traveling direction of the shearer, the standby pressure of the hydraulic system, the height of the hydraulic oil in the pump box, the cooling water flow rate, pressure, tank temperature, and the left and right drum heights.

[0029] b. The operation parameters of the support include the support column pressure, the pushing stroke, the control mode, the emergency stop state of the support controller, the communication state, the communication state between the driver and the support controller, the progress of the working face, including the current - shift and cumulative progress, the single - frame single - action, the group - pushing scraper conveyor, the group - extending and retracting the rib protection, and the action coding data of the group - extending and retracting the telescopic beam.

[0030] c. The operation parameters of the crusher - conveyor - scraper conveyor include the temperature, pressure, flow rate, displacement, rotational speed, switch - state display, loop operation state, current value, voltage value, leakage, open - phase, and overload data of the equipment reducer and motor.

[0031] d. The operation parameters of the pump station include the pump - station outlet pressure, the pump - station oil temperature, the pump - station oil - level state, the action condition of the pump - station solenoid valve, the liquid - tank liquid level, and the emulsion - tank oil - level data.

[0032] Example 5 For the different level divisions in S5, the higher the level, the more serious the potential danger. Among them: I: Pop-up warning on the PC side of the big data intelligent mining platform, no processing; II: Pop-up warning - audible and visual warning of the underground lighting warning device - audible and visual prompts in the ground dispatching room and underground voice broadcast; III: Pop-up warning - audible and visual warning - voice broadcast prompt - one-key start and stop of fully-mechanized mining equipment.

[0033] System pop-up warning: The warning information appears in the form of a pop-up window on the system interface, and the pop-up window duration is at least 5 seconds to ensure that the operator can see and notice the warning information.

[0034] Audible and visual warning: The system emits sound and light signals. The sound loudness reaches at least 60 decibels, and the light signal flashing frequency is at least 2 times per second to attract the attention of the operator.

[0035] Voice broadcast prompt: The system broadcasts the warning information through voice synthesis technology. The broadcast content is clear and the speech rate is moderate to ensure that the operator can accurately understand the warning information.

[0036] One-key start and stop of the shearer: After the warning information is confirmed, the operator can perform emergency stop or start operations on the shearer through the one-key start and stop button on the system interface. The stop operation should be carried out after confirming safety to avoid damage to equipment and personnel. The start operation should be carried out after troubleshooting and ensuring safety.

[0037] Example 6 For S4 and S5, different control instructions of different levels are issued corresponding to different abnormal situations as follows: a. When the AI intelligent recognition module recognizes that there is a foreign object on the belt, when the recognition result of the foreign object is a non-metallic sharp object such as gangue and does not affect the normal operation of the belt, a pop-up warning is given, no processing is done, and a level I control instruction is issued; when it is judged that there is a foreign object, when the recognition result of the foreign object is a sharp metal object such as a bolt, angle iron, I-beam, or anchor bolt that affects the normal operation of the belt and even causes belt damage, the equipment control system can automatically give an audible and visual alarm, voice prompt, and even perform one-key start and stop operations on the shearer; the foreign object recognition process is as Figure 2 shown, mainly including three parts: dataset construction, foreign object recognition model establishment, and foreign object AI intelligent recognition. First, the collected dataset is preprocessed clearly, and the preprocessed dataset is labeled and a dataset is prepared; then, the entire data integration ratio is randomly divided into a training set and a test set, and the training set data is imported into the deep learning network model for training to obtain a foreign object recognition model; finally, the foreign object recognition model is used to monitor the video in real time and obtain the foreign object recognition result.

[0038] b. When the AI intelligent recognition module identifies that the operating parameters of the shearer, support, crusher - conveyor - scraper conveyor, and pump station are abnormal (the current and voltage loads exceed the rated current and voltage, and the voltage and current data are abnormal), and the operating conditions are poor, affecting the safe and efficient mining of the working face. When the system identifies one or more of the above abnormal parameters, it issues a level III warning, and the system pops up a warning - audible and visual alarm - voice broadcast prompt - one - key start and stop of the shearer.

[0039] Abnormal operating parameters of the shearer: Abnormal current: Measured current value > rated current value (for example: measured current is 150A, while the rated current is 120A); Abnormal voltage: Measured voltage value > rated voltage value or measured voltage value < allowable minimum voltage value (for example: measured voltage is 420V, while the rated voltage is 380V; or measured voltage is 340V, lower than the allowable minimum voltage of 360V); Abnormal load: Load rate > rated load rate (for example: load rate is 120%, while the rated load rate is 100%).

[0040] Abnormal operating parameters of the support: Abnormal hydraulic pressure: Measured hydraulic pressure value > maximum allowable pressure value or measured hydraulic pressure value < minimum allowable pressure value (specific values need to be set according to the actual equipment); Abnormal support moving speed: Measured moving speed > maximum allowable speed or measured moving speed < minimum allowable speed (specific values need to be set according to the actual equipment); Abnormal operating parameters of the crusher - conveyor - scraper conveyor: Abnormal current: The same as above (quantified separately for the crusher, conveyor, and scraper conveyor); Abnormal voltage: The same as above (quantified separately for the crusher, conveyor, and scraper conveyor); Abnormal vibration: Vibration amplitude > allowable maximum vibration amplitude (specific values need to be set according to the actual equipment); Abnormal rotational speed: Measured rotational speed < allowable minimum rotational speed or measured rotational speed > allowable maximum rotational speed (specific values need to be set according to the actual equipment).

[0041] Abnormal operating parameters of the pump station: Abnormal flow: Measured flow < allowable minimum flow or measured flow > allowable maximum flow (specific values need to be set according to the actual equipment); Abnormal pressure: Measured pressure value > maximum allowable pressure value or measured pressure value < minimum allowable pressure value (specific values need to be set according to the actual equipment); Abnormal oil temperature: Oil temperature > allowable maximum oil temperature (for example: oil temperature reaches 85°C, while the allowable maximum oil temperature is 80°C).

[0042] c. The real-time video data of the shearer following the operation in the fully mechanized coal mining face captured by the hydraulic support camera. When the AI intelligent recognition module recognizes that the hydraulic support fails to move in time (after the shearer cuts coal, the 3 supports behind the rear drum have not moved and supported in time) or the rib protection plate fails to retract in time (the rib protection plates of the 3 supports in front of the front drum of the shearer have not retracted in time, posing a risk of collision between the shearer and the support), which affects the normal mining of the working face, a level III warning is issued and the shearer is started and stopped with one key. d. When the AI intelligent recognition module recognizes that there are obvious cracks or fissures in the scraper chain of the scraper conveyor, and there is a risk of breakage at any time, which affects the normal operation of the scraper conveyor, an equipment danger warning is issued, the shearer is started and stopped with one key, and a level III control instruction is issued. Using the cameras arranged at the head and tail of the scraper conveyor to collect the images of the chain plates. After preprocessing the collected chain plate images, they enter the defect monitoring process. The defect monitoring of the chain plates adopts secondary template matching defect monitoring, and both monitoring processes use multi-template matching. Using the constructed non-destructive chain plate images for the first template matching defect monitoring to obtain suspected defect images; storing the suspected defect images and entering the second template matching defect monitoring. Using the constructed crack and broken chain defect reference template library to conduct secondary monitoring of the suspected defect images. After comprehensive judgment, the AI intelligent recognition module displays the final result. When it recognizes that there are large cracks and broken chain situations during the operation of the scraper, it links the equipment control system, issues a level III control instruction, gives an audible and visual alarm, and intelligently shuts down the shearer and the scraper conveyor.

[0043] e. Using the real-time video data of the shearer following the operation in the fully mechanized coal mining face captured by the hydraulic support camera. When the AI intelligent recognition module recognizes that there are people breaking into the dangerous area of the fully mechanized coal mining face, a pop-up warning, an audible and visual alarm, and a voice prompt are given until the people leave this area, and a level II control instruction is issued. Hydraulic support camera: Video resolution: High definition (e.g., 1920x1080 pixels) to ensure clear video images; Video frame rate: At least 30 frames per second to ensure smooth video; Shooting range: Cover the key areas of the fully mechanized coal mining face to ensure no blind spots; Real-time performance: The video data is transmitted to the AI intelligent recognition module in real time with a delay of no more than 0.5 seconds.

[0044] AI intelligent recognition module: Recognition accuracy: The recognition accuracy for people breaking into the dangerous area of the fully mechanized coal mining face is not less than 95%; Recognition speed: The time from video data input to recognition result output is not more than 1 second; Recognition range: According to the shooting range of the hydraulic support camera, the AI intelligent recognition module can accurately recognize all people within this range.

[0045] Personnel departure detection: The AI intelligent recognition module continuously monitors the dangerous area. When it recognizes that the personnel have completely left the area, the system automatically cancels the early warning and control instructions.

[0046] The detection accuracy rate of personnel departure is not less than 98%, ensuring that there will be no unnecessary shutdown or delay due to misjudgment.

[0047] f. When the methane concentration exceeds 1.00%, the carbon monoxide concentration exceeds 24 ppm, and the dust concentration exceeds 1.5 mg / m 3 , when the sensors detect that the indicators in the stope range exceed the limits, a pop-up warning will be immediately issued on the platform, accompanied by audible and visual alarms and voice prompts underground, one-key start and stop of the fully-mechanized mining equipment underground, the dispatching room will direct the underground personnel to evacuate in an orderly manner, and a Class III control instruction will be issued; g. When the number of personnel at the working face reaches the upper limit (underground workers wear positioning sensors themselves, and when the number of personnel in the working face area exceeds the limit, the system will automatically give an early warning), a voice prompt will be issued. When subsequent personnel enter the range of the working face area, a pop-up warning will be issued on the big data platform, a Class II early warning will be issued, accompanied by audible and visual warnings and voice alarms underground until the extra personnel in the working face evacuate from this area.

[0048] The above method aims to reduce the design difficulty, construct an effective equipment linkage control strategy, realize the real-time supervision of people's unsafe behaviors and the equipment linkage control, so as to ensure the safe and efficient mining of the fully-mechanized mining face and the high-efficiency and stable operation of the equipment. At the same time, the introduction of the intelligent mining interlock control system simplifies the design of the fault-tolerant system while significantly improving the ability to identify potential hidden dangers in the on-site process execution and the reaction and response ability to emergencies, effectively reducing the accident rate and providing a strong guarantee for the safe and efficient mining of the fully-mechanized mining face.

[0049] The present invention proposes an interlock control method for fully mechanized mining equipment based on AI recognition and sensor fusion feedback. The core lies in promoting the in-depth integration and development of coal mine AI intelligent technology and safety production. By integrating multiple key systems such as real-time monitoring of the production system, evaluation of the process execution degree, monitoring of the equipment safety status, consideration of environmental factors, monitoring of personnel safety behaviors, and existing disaster system videos, the present invention utilizes AI intelligent recognition technology and AR augmented reality technology, combined with the real-time working condition information feedback of the production site, to comprehensively evaluate the process standardization degree, equipment working condition safety parameters, personnel safety behaviors, and the integrity of human-computer interaction process execution of the current production scenario. On this basis, an effective combined control strategy is formed, and through intelligent control collaborative services, a series of protection measures such as audible and visual warnings and control interlock protection are carried out on the equipment to respond to complex production safety emergency requirements in an efficient, fast, and flexible manner. It can also, according to the on-site process requirements, make full use of rich on-site production control experience and process accumulation, link various production systems and equipment, form an efficient interlock control strategy, and carry out linkage and collaborative control on the major production systems of the mine. This innovation not only improves the intelligent mining level of the mine but also further meets the actual production requirements on site, providing a new solution for the safety production and intelligent development of the coal mine industry.

[0050] Example 7 The following Table 1 is a set of data illustrating the specific application effects of the interlock control method for fully mechanized mining equipment based on AI recognition and sensor fusion feedback: Table 1

[0051] In this set of data, the data of the fully mechanized mining operation site are collected in real time by sensors, and the AI recognition system processes and analyzes the collected data and issues corresponding control instructions according to the recognition results. At 10:22, the camera detects that a person is approaching the shearer, and the system immediately issues an audible and visual warning and decelerates until the person evacuates this area; at 12:06, the infrared sensor detects that the temperature of the shearer motor is too high, and the system triggers the cooling system and adjusts the load to prevent overheating; at 15:41, the lidar detects an obstacle ahead, and the system brakes urgently and plans an obstacle avoidance path to ensure safety; at 10:03, the pressure sensor detects abnormal pressure of the hydraulic support, and the system adjusts the supporting force to maintain balance. The issuance of these control instructions is based on the results of AI recognition and sensor fusion feedback, realizing the interlock control and intelligent management of fully mechanized mining equipment.

Claims

1. An interlock control method for fully-mechanized mining equipment based on AI recognition and sensor fusion feedback, characterized in that: Specifically, it includes the following steps: Step 1: Real-time collect the video images of underground coal mining and transportation, and store and transmit the obtained video stream to the AI intelligent recognition module; Step 2: Install sensors underground to monitor the operating conditions of fully-mechanized mining equipment and environmental indicators in the stope space; Step 3: Connect the monitoring data in the AI intelligent recognition module in Step 1 and the sensors in Step 2 to the big data platform; Step 4: Use the AI intelligent recognition module to identify abnormal situations at the underground operation site; Step 5: Based on the abnormal situations identified in Step 4 and the monitoring data of the sensors, issue control instructions with different warning levels to the control system of the fully-mechanized mining equipment; Step 6: According to the different-level control instructions issued in Step 5, conduct linkage control on the fully-mechanized mining equipment to implement the interlocking control protection mechanism of the fully-mechanized mining equipment.

2. The integrated mining equipment interlock control method based on AI recognition and sensor fusion feedback according to claim 1, characterized in that: In Step 1, install cameras underground to obtain the video images of underground coal mining and coal transportation.

3. The integrated mining equipment interlock control method based on AI recognition and sensor fusion feedback according to claim 1, wherein: The sensors in Step 2 include carbon monoxide sensors, methane sensors, and dust sensors.

4. The integrated mining equipment interlock control method based on AI recognition and sensor fusion feedback according to claim 1, characterized in that: The warning levels in Step 5 include: Level I: Pop-up warning on the PC side of the big data intelligent mining platform, no processing; Level II: Pop-up warning - audible and visual alarm of the underground lighting alarm device - audible and visual prompt in the ground dispatching room and underground; Level III: Pop-up warning - audible and visual alarm - audible and visual prompt - one-key start and stop of the fully-mechanized mining equipment.

5. The integrated mining equipment interlock control method based on AI recognition and sensor fusion feedback according to claim 4, wherein: The specific process of Step 5 is as follows: Step 5.1: When the AI intelligent recognition module recognizes that there is a foreign object on the belt and it does not affect the normal operation of the belt, a pop-up warning is issued, no processing is done, and a Level I control instruction is issued; when the AI intelligent recognition module recognizes that there is a foreign object on the belt and it affects the normal operation of the belt or even causes belt damage, the equipment control system can automatically give an audible and visual alarm, voice prompt, issue a Level III control instruction, and implement one-key start and stop of the shearer; Step 5.2: If the AI intelligent recognition module recognizes that the operating parameters of the shearer, the operating parameters of the support, the operating parameters of the crusher - conveyor - scraper conveyor, and the operating parameters of the pump station are abnormal and affect the normal operation of the fully-mechanized mining equipment, a Level III warning is issued, and the system gives a pop-up warning - audible and visual alarm - voice prompt - one-key start and stop of the shearer; Step 5.3: When the AI intelligent recognition module recognizes that the hydraulic support has problems such as untimely shield movement and untimely retraction of the rib protection plate, which affect the normal mining of the working face, a Level III warning is issued, and one-key start and stop of the shearer is implemented; Step 5.4: When the AI intelligent recognition module recognizes that the scraper chain of the scraper conveyor has cracks and affects the normal operation of the scraper conveyor, an equipment danger warning is given, one-key start and stop operation of the shearer is performed, and a Level III control instruction is issued; Step 5.5: Utilize the real-time video data of the shearer following the shearer operation in the fully-mechanized mining face captured by the hydraulic support camera. When the AI intelligent recognition module recognizes that a person has entered the dangerous area of the fully-mechanized mining face, a pop-up warning, audible and visual alarm, and voice prompt are given until the person leaves this area, and a Level II control instruction is issued; Step 5.6, when the sensor monitors that the methane concentration exceeds 1.00%, the carbon monoxide concentration exceeds 24 ppm, and the dust concentration exceeds 1.5 mg / m 3 ³, immediately issue a pop-up warning on the platform, give an audible and visual alarm underground, and voice prompt. For the fully-mechanized mining equipment underground, start and stop it with one key. The dispatching room commands the underground personnel to evacuate in an orderly manner and issues a Class III control instruction; Step 5.7, when the number of personnel on the working face reaches the upper limit of personnel, voice prompts are given. If there are subsequent personnel entering the working face area, the big data platform will pop up an alarm, issue a level-II early warning, and give audible and visual warnings underground and voice alarms until the redundant personnel on the working face evacuate this area.

6. The interlocking control method for fully-mechanized mining equipment based on AI recognition and sensor fusion feedback according to claim 5, wherein: In the said step 5.2, the abnormal operation parameters of the shearer include: Abnormal current: measured current value > rated current value; Abnormal voltage: measured voltage value > rated voltage value or measured voltage value < allowable minimum voltage value; Abnormal load: load rate > rated load rate.

7. The integrated mining equipment interlock control method based on AI recognition and sensor fusion feedback according to claim 6, wherein: In the said step 5.2, the abnormal operation parameters of the support include: Abnormal hydraulic pressure: measured hydraulic pressure value > maximum allowable pressure value or measured hydraulic pressure value < minimum allowable pressure value; Abnormal support moving speed: measured moving speed > maximum allowable speed or measured moving speed < minimum allowable speed.

8. The integrated mining equipment interlock control method based on AI recognition and sensor fusion feedback according to claim 6, characterized in that: In the said step 5.2, the abnormal operation parameters of the crusher - conveyor - scraper conveyor include: Abnormal vibration: vibration amplitude > allowable maximum vibration amplitude; Abnormal rotation speed: measured rotation speed < allowable minimum rotation speed or measured rotation speed > allowable maximum rotation speed.

Images