Power plant production area abnormal behavior analysis method and system

Abstract

The invention relates to the technical field of power production safety management, in particular to a power plant production area abnormal behavior analysis method and system. Comprising a data acquisition unit, a 5G network management unit, a data management unit, a personnel management unit, an equipment management unit, a behavior analysis unit, a production area abnormity management unit and an abnormity alarm linkage management unit. The data acquisition unit multiplexes existing digital cameras, face recognition access control equipment and a face snapshot machine in a factory area, the 5G network management unit is used for constructing a 5G local area network for a power plant, and the data management unit is used for data storage and playback; the personnel management unit is used for managing factory local employees, outsourcing employees and visitors; the behavior analysis unit is used for deploying a multi-dimensional abnormal behavior analysis model, the production area abnormity management unit is used for receiving analysis data sent by the behavior analysis unit, and the abnormity alarm linkage management unit broadcasts abnormity of a certain person or certain equipment; the complex environment adaptability is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power production safety management, and particularly to a method and system for analyzing abnormal behaviors in a power plant production area. Background Art

[0002] In addition to formal employees in the power plant area, there are also many outsourced personnel and temporary visitors. The number of outsourced personnel and visitors entering and leaving the plant area every day reaches hundreds. Coupled with the large production site in the plant area, scattered personnel distribution, large differences in the quality of operating personnel, and a working environment with high temperature, high pressure, ultra-high voltage, dust, high-altitude falls, high-altitude falling objects, acids and alkalis, harmful gases, and mechanical injuries, there are many various potential safety hazards. Operating personnel cannot accurately control their activity areas when entering the plant. Once a personnel accident occurs due to unfamiliarity with the plant environment and entering a high-risk area, it will have a bad impact. There are high-risk factors such as high temperature, high pressure, and harmful gases, and the personnel flow is large (including formal employees, outsourced personnel, and visitors). There are three current personnel abnormal behavior management technologies; the first technology is relatively traditional and is adopted by most power plants, that is, through the video monitoring equipment in the plant area, monitored by a special person in the control center, and those who violate the regulations are immediately notified to themselves; the second is the manual spot-check method, where inspectors conduct inspections irregularly and impose fines if they are caught; the third is the video analysis technology, which analyzes video images and alarms for some violations; the traditional safety management means have the following problems:

[0003] 1. Video monitoring relies on manual inspections: The monitoring personnel are prone to fatigue, resulting in missed inspections;

[0004] 2. The efficiency of manual spot-checks is low: It cannot cover the entire area, and the missed inspection rate is high;

[0005] 3. The existing video analysis technology has a high false alarm rate: The algorithm has poor adaptability and is difficult to meet the requirements of complex scenarios in power plants. Summary of the Invention

[0006] To solve the above technical problems, the present invention provides a method and system for analyzing abnormal behaviors in a power plant production area.

[0007] An abnormal behavior analysis system for a power plant production area according to the present invention includes a data acquisition unit, a 5G network management unit, a data management unit, a personnel management unit, an equipment management unit, a behavior analysis unit, a production area abnormality management unit, and an abnormal alarm linkage management unit;

[0008] The data acquisition unit reuses the existing digital cameras, face recognition access control devices, and face capture cameras in the plant area, and adds intelligent cameras supporting 5G (including infrared sensing modules) to collect video, personnel location, and equipment status data in real time and send them;

[0009] The 5G network management unit is used to build a 5G local area network for the power plant, support low-latency (<50ms) transmission of video streams and interconnection of multiple devices, realize high-speed transmission of video data and device interconnection, and ensure real-time and clear video images;

[0010] The data management unit uses the 5G local area network to receive data such as videos, personnel positions, and equipment status sent by the data acquisition unit, and performs data storage and playback;

[0011] The personnel management unit is used to manage the regular employees, outsourced employees, and visitors in the factory area;

[0012] The equipment management unit, with the support of the 5G local area network, conducts multiple management of the factory area equipment, including Internet of Things, basic information, and location-based management;

[0013] The behavior analysis unit deploys a multi-dimensional abnormal behavior analysis model, including multi-dimensional analysis units such as crossing the boundary, smoking, not wearing a safety helmet, falling, safety, fire protection, and equipment failure;

[0014] The production area anomaly management unit is used to receive the analysis data sent by the behavior analysis unit. When it finds that the analysis result is abnormal, such as an alarm, it sends an electrical signal to the anomaly alarm linkage management unit

[0015] The anomaly alarm linkage management unit is used to receive the electrical signal sent by the production area anomaly management unit, realize local broadcast linkage, and announce the anomaly of someone or a certain device; The data acquisition unit reuses existing devices (cameras, access control) and adds intelligent cameras to realize multi-dimensional acquisition of video, infrared sensing, and Internet of Things device data, covering the full scenarios of personnel behavior, environmental safety, and equipment operation. Through the 5G network management unit, a low-latency (<50ms) transmission channel is built to ensure real-time synchronization of video streams, personnel positions, and equipment status data, support second-level anomaly response (such as the crossing boundary alarm trigger time <1 second), and the behavior analysis unit integrates 7 types of anomaly detection algorithms such as crossing the boundary, smoking, and not wearing a safety helmet, and combines infrared sensing and equipment status data for cross-verification, reducing the false alarm rate from >20% in the traditional solution to <5%; The personnel management unit and the equipment management unit are integrated into the same system, reducing the manpower and time consumption of multi-platform switching. Through the production area anomaly management unit, the risk level is dynamically divided (such as first-level high-risk alarm, second-level potential risk), reducing the interference of ineffective alarms, realizing high timeliness, multi-source data fusion, intelligent grading judgment, automated closed-loop management, resource service reuse and cost savings, and improving the adaptability to complex environments.

[0016] Preferably, the crossing boundary analysis unit, based on personnel management and video analysis and combined with the electronic fence and personnel positioning data, determines in real time whether to enter a high-risk area, and pushes the analysis result to the production area anomaly management. If there is an alarm, the anomaly alarm linkage management is started;

[0017] Smoking analysis unit, which conducts out-of-bounds analysis based on personnel management and video analysis, and pushes the analysis results to the abnormal management of the production area. If there is an alarm, it activates the abnormal alarm linkage management;

[0018] Helmetless analysis unit, which conducts helmetless analysis based on personnel management and video analysis, and pushes the analysis results to the abnormal management unit of the production area. If there is an alarm, it activates the abnormal alarm linkage management;

[0019] Safety analysis unit, which is used for the analysis of personnel falls, fights, climbs, and equipment failures, and pushes the analysis results to the abnormal management unit of the production area. If there is an alarm, it activates the abnormal alarm linkage management;

[0020] Fire analysis unit, which is used for the analysis of equipment leaks, fires, etc. in the power plant area, and pushes the analysis results to the abnormal management unit of the production area. If there is an alarm, it activates the abnormal alarm linkage management; The out-of-bounds analysis unit determines in real time whether a high-risk area has been entered based on personnel management, video analysis, and in combination with electronic fence and personnel positioning data. The smoking analysis unit conducts out-of-bounds analysis based on personnel management and video analysis. The helmetless analysis unit conducts helmetless analysis based on personnel management and video analysis. The safety analysis unit analyzes personnel falls, fights, climbs, and equipment failures. The fire analysis unit analyzes equipment leaks, fires, etc. in the power plant area. The above units all push the analysis results to the abnormal management unit of the production area. If there is an alarm, it activates the abnormal alarm linkage management to achieve multi-source data fusion and intelligent hierarchical judgment.

[0021] Preferably, it further includes a drone inspection unit. The drone inspection unit includes a drone main body, propellers, support feet and a housing. Multiple groups of propellers are arranged on the drone main body. Two groups of support feet are installed at the bottom end of the drone main body. A housing is arranged at the bottom end of the drone main body. A square opening is arranged at the bottom end of the housing. It further includes a sealing component, a detection component, a transmission component and a gas analysis component. A circular hole is installed at each of the left and right ends of the housing. A sealing component is arranged on the housing. The sealing component seals the square opening. The detection component is installed inside the housing. A transmission component is arranged between the detection component and the sealing component. Two gas analysis components are installed inside the housing; when it is necessary to inspect the hidden corners of the power plant, the staff operates multiple groups of propellers to rotate quickly, so that the drone main body flies in the air. The detection component is operated to extend from the inner side of the housing to the lower part of the housing. While the detection component extends downward, the sealing component is released from sealing the square opening through the transmission component, so that the detection component extends to the lower part of the housing through the square opening. At the same time, while the sealing component releases the seal of the square opening, the sealing component makes the suction ends of the two gas analysis components respectively pass through a circular hole and extend to the outside of the housing, so that the gas analysis components absorb and analyze the external gas, improving the convenience. After use, while the detection component enters the housing, the sealing component seals the square opening. At the same time, the sealing component makes the suction ends of the gas analysis components re-enter the housing, improving the convenience.

[0022] Preferably, the sealing component includes support plates, optical bars, springs, sealing plates, insertion rods and square frames. Two pairs of support plates are arranged at the bottom end of the housing. A group of optical bars are respectively arranged between each group of support plates. The two sealing plates are respectively slidably connected to the two optical bars. Two springs are respectively sleeved on the outer sides of each group of optical bars. The two sealing plates are both between the two springs. A group of insertion rods are respectively arranged at the two side ends of one sealing plate. A group of square frames are respectively installed at the two side ends of the other sealing plate. The insertion rods are adapted to the square frames; in the normal state, the two sealing plates are mirror-symmetrically close to each other under the action of the elastic force of the corresponding springs. The two sealing plates cooperate with each other to seal the square opening. When the detection component needs to extend to the lower part of the housing through the square opening, the detection component pulls the two sealing plates through the two transmission components, so that the two sealing plates are mirror-symmetrically away from each other under the cooperation of the two optical bars, so that the two sealing plates release the seal of the square opening.

[0023] Preferably, the detection component includes a servo electric cylinder, a lifting plate, an L-shaped guide plate, a smooth rod, a high-definition camera, and an infrared imager. A servo electric cylinder is installed at the top end inside the housing. The bottom mobile end of the servo electric cylinder is installed with a lifting plate. Two groups of L-shaped guide plates are arranged at the top end of the lifting plate. Two groups of smooth rods are fixedly installed inside the housing. One group of L-shaped guide plates are respectively slidably connected with one group of smooth rods. The bottom end of the lifting plate is respectively provided with a high-definition camera and an infrared imager. When it is necessary to conduct a patrol inspection of the power plant, the staff operates the servo electric cylinder to extend, so that the lifting plate moves downward under the guiding cooperation of the L-shaped guide plate and the smooth rod. The lifting plate pulls the two sealing plates through two transmission components, so that the two sealing plates move away mirror-symmetrically. The lifting plate drives the high-definition camera and the infrared imager to extend below the housing through the square opening, improving convenience. After use, when the high-definition camera and the infrared imager enter the housing, the two sealing plates seal the square opening, improving safety protection.

[0024] Preferably, the transmission component includes a first bracket, a first fixed pulley, a second bracket, a second fixed pulley, and a rope. Two groups of first brackets are arranged at the bottom end of the housing. One group of first fixed pulleys are respectively rotatably arranged on each group of first brackets. Two groups of second brackets are arranged at the top end inside the housing. One group of second fixed pulleys are respectively rotatably arranged on each group of second brackets. The outer ends of the two sealing plates are respectively connected with the first ends of a group of ropes. The second ends of the two groups of ropes respectively bypass the bottom ends of one group of first fixed pulleys and enter the housing and bypass the top ends of the second fixed pulleys and are connected with the top end of the lifting plate. When in use, the servo electric cylinder extends, so that the lifting plate moves downward. The lifting plate makes the two groups of ropes pull the two sealing plates synchronously under the guiding cooperation of the second fixed pulley and the first fixed pulley, so that the two sealing plates perform a mirror-symmetric separation operation. When the lifting plate enters the housing, the two sealing plates move closer mirror-symmetrically, so that the two sealing plates seal the square opening again.

[0025] Preferably, the gas analysis component includes a gas analyzer main body, a gas transmission hose, a cylinder body, a suction head, a filter screen, a moving plate, and a support rod. Two groups of gas analyzer main bodies are installed at the rear end inside the housing. One group of gas transmission hoses is respectively arranged at the input end of each group of gas analyzer main bodies. The input end of each group of gas transmission hoses is communicated with the output end of a cylinder body. One group of moving plates is respectively arranged at the top end of each group of sealing plates. One group of support rods is respectively installed on each group of moving plates. The other end of each group of support rods is respectively installed with a cylinder body. The input end of the cylinder body is provided with a suction head, and the input end of the suction head is installed with a filter screen. Notch openings are respectively arranged on the left and right sides of the square opening; when the two groups of sealing plates move away from each other in a mirror image manner, the sealing plates respectively drive a cylinder body to move away from each other in a mirror image manner through the support rods, so that the suction heads on each group of cylinder bodies extend through the circular holes to the outside of the housing. Start the gas analyzer main body, so that the suction head inhales the external gas and analyzes it through the gas analyzer main body. After use, when the two groups of sealing plates re-seal the square opening, the two suction heads re-enter the housing through the circular holes for safety protection.

[0026] Preferably, it further includes a fixing plate, a rotating shaft, and a cover plate. Two groups of fixing plates are respectively arranged at the left and right ends of the housing. A rotating shaft is rotatably arranged between the two groups of fixing plates. One group of cover plates is respectively arranged on each group of rotating shafts. A torsion spring is arranged on the rotating shaft, and the torsion spring makes the cover plate seal the circular hole; in the normal state, the cover plate seals the circular hole with the cooperation of the torsion spring. During use, the two groups of support rods drive the two suction heads to move away from each other in a mirror image manner, so that the two suction heads push open the corresponding cover plates for sampling. After use, the two suction heads re-enter the housing, and the two cover plates re-seal the circular hole.

[0027] Preferably, buffer pads are respectively arranged at the ends of the two groups of sealing plates close to each other; the buffer pads avoid the hard collision when the two groups of sealing plates seal the square opening, and improve the service life.

[0028] A method for analyzing abnormal behaviors in the production area of a power plant includes the following steps:

[0029] Step 1: Real-time collect video streams, face recognition data, and device sensor data through the 5G network;

[0030] Step 2: Use multi-threaded parallel processing technology to classify and analyze the data: Extract frames from the video stream, input the behavior recognition model, output the probabilities of behaviors such as smoking and not wearing a safety helmet, combine the coordinates of the electronic fence and the personnel positioning data, and trigger an out-of-bounds alarm; Analyze the device sensor data (such as temperature, vibration) to determine whether there is a fault or leakage;

[0031] Step 3: Alarm at different levels according to the analysis results:

[0032] Level 1 alarm (high-risk behavior, such as unauthorized entry into the high-voltage area): Immediately trigger a broadcast alarm and notify the safety officer;

[0033] Secondary alarm (potential risks, such as not wearing a safety helmet): Push to the management platform and record;

[0034] Step 4: Store all data and generate statistical reports, supporting retrospective analysis by time, region, and behavior type.

[0035] Compared with the prior art, the beneficial effects of the present invention are as follows: The data acquisition unit reuses existing devices (cameras, access control) and adds intelligent cameras to achieve multi-dimensional acquisition of video, infrared sensing, and Internet of Things device data, covering all scenarios of personnel behavior, environmental safety, and equipment operation. A low-latency (<50 ms) transmission channel is built through the 5G network management unit to ensure real-time synchronization of video streams, personnel positions, and device status data, supporting second-level anomaly response (such as the intrusion alarm trigger time <1 second). The behavior analysis unit integrates 7 types of anomaly detection algorithms such as intrusion, smoking, and not wearing a safety helmet, and combines infrared sensing and device status data for cross-verification, reducing the false alarm rate from >20% in the traditional solution to <5%; The personnel management unit and the device management unit are integrated into the same system, reducing the manpower and time consumption of multi-platform switching. The risk level is dynamically divided through the production area anomaly management unit (such as first-level high-risk alarm, second-level potential risk), reducing the interference of invalid alarms, achieving high timeliness, multi-source data fusion, intelligent hierarchical judgment, automated closed-loop management, resource service reuse, and cost savings, and improving the adaptability to complex environments. Brief Description of the Drawings

[0036] Figure 1 is the system block diagram of the present invention;

[0037] Figure 2 is the axonometric schematic diagram of the drone inspection unit of the present invention;

[0038] Figure 3 is the exploded structure schematic diagram of the drone main body and the housing, etc.;

[0039] Figure 4 is the enlarged structure schematic diagram of the servo electric cylinder and the rope, etc.;

[0040] Figure 5 is Figure 4 the partial enlarged structure schematic diagram of part A in;

[0041] Figure 6 is the enlarged structure schematic diagram of the sealing plate and the cover plate, etc.;

[0042] Figure 7 is Figure 6 the partial enlarged structure schematic diagram of part B in;

[0043] Figure 8 is Figure 6 the partial enlarged structure schematic diagram of part C in;

[0044] Figure 9 It is an enlarged structural schematic diagram of structures such as the polished rod and the main body of the gas analyzer;

[0045] Figure 10 It is an enlarged structural schematic diagram of structures such as the servo electric cylinder and the high-definition camera;

[0046] Figure 11 It is an enlarged structural schematic diagram of structures such as the main body of the gas analyzer and the cylinder body.

[0047] Markings in the attached drawings: 101, UAV main body; 102, propeller; 103, support feet; 104, housing; 201, support plate; 202, polished rod; 203, spring; 204, sealing plate; 205, insertion rod; 206, square box; 301, servo electric cylinder; 302, lifting plate; 303, L-shaped guide plate; 304, polished rod; 305, high-definition camera; 306, infrared imager; 401, first bracket; 402, first fixed pulley; 403, second bracket; 404, second fixed pulley; 405, rope; 501, main body of the gas analyzer; 502, gas transmission hose; 503, cylinder body; 504, suction head; 505, filter screen; 506, moving plate; 507, support rod; 601, fixing plate; 602, rotating shaft; 603, cover plate. Detailed implementation manners

[0048] To facilitate the understanding of the present invention, the present invention will be described more comprehensively below with reference to the relevant attached drawings. The present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, these embodiments are provided to make the disclosure of the present invention more thorough and comprehensive.

[0049] Embodiment 1

[0050] As Figures 1 to 11 shown, an abnormal behavior analysis system for the power plant production area of the present invention includes a data acquisition unit, a 5G network management unit, a data management unit, a personnel management unit, an equipment management unit, a behavior analysis unit, a production area abnormal management unit, and an abnormal alarm linkage management unit;

[0051] The data acquisition unit reuses the existing digital cameras, face recognition access control devices, and face capture cameras in the plant area, and newly adds intelligent cameras (including infrared sensing modules) supporting 5G, which are used to collect video, personnel positions, and equipment status data in real time and send them;

[0052] The 5G network management unit is used to build a 5G local area network for the power plant, support the transmission of video streams with low latency (<50 ms) and the interconnection of multiple devices, realize the high-speed transmission of video data and the interconnection of devices, and ensure that the video images are real-time and clear;

[0053] The data management unit uses the 5G local area network to receive the data sent by the data collection unit, including video, personnel location and equipment status, and performs data storage and playback;

[0054] Personnel management unit, used to manage the factory's own employees, outsourced employees and visitors;

[0055] The equipment management unit, with the support of 5G local area network, conducts multiple management of factory equipment based on IoT, basic information and location;

[0056] Behavior analysis unit, deploying multi-dimensional abnormal behavior analysis models, including crossing the boundary, smoking, not wearing a helmet, falling, safety, fire fighting and equipment failure;

[0057] The production area abnormality management unit is used to receive the analysis data sent by the behavior analysis unit. If the analysis results are abnormal, if there is an alarm, an electrical signal is sent to the abnormal alarm linkage management unit.

[0058] The abnormal alarm linkage management unit is used to receive the electrical signal sent by the abnormal management unit in the production area, realize local broadcast linkage, and report the abnormality of a person or a device;

[0059] The out-of-bounds analysis unit, based on personnel management and video analysis and combined with electronic fence and personnel positioning data, determines in real time whether a person has entered a high-risk area, and pushes the analysis results to the production area abnormal management. If there is an alarm, the abnormal alarm linkage management is initiated;

[0060] The smoking analysis unit conducts cross-border analysis based on personnel management and video analysis, and pushes the analysis results to the abnormal management of the production area. If there is an alarm, the abnormal alarm linkage management is initiated;

[0061] The helmet-not-wearing analysis unit performs helmet-not-wearing analysis based on personnel management and video analysis, and pushes the analysis results to the production area abnormality management unit. If there is an alarm, the abnormal alarm linkage management is started;

[0062] The safety analysis unit is used to analyze personnel falls, fights, climbing, and equipment failures, and push the analysis results to the abnormal management unit in the production area. If there is an alarm, the abnormal alarm linkage management is started;

[0063] The fire analysis unit is used to analyze equipment leakage, fire, etc. in the power plant area, and push the analysis results to the abnormal management unit in the production area. If there is an alarm, the abnormal alarm linkage management is initiated.

[0064] In this embodiment, the data acquisition unit reuses existing devices (cameras, access control) and adds intelligent cameras to achieve multi-dimensional acquisition of video, infrared sensing, and IoT device data, covering all scenarios of personnel behavior, environmental safety, and equipment operation. A low-latency (<50 ms) transmission channel is built through the 5G network management unit to ensure real-time synchronization of video streams, personnel positions, and equipment status data, supporting second-level anomaly response (such as the intrusion alarm trigger time <1 second). The behavior analysis unit integrates 7 types of anomaly detection algorithms such as intrusion, smoking, and not wearing a safety helmet, and combines infrared sensing with equipment status data for cross-verification, reducing the false alarm rate from >20% in the traditional solution to <5%; the personnel management unit and the equipment management unit are integrated into the same system, reducing the manpower and time consumption of multi-platform switching.

[0065] The risk level is dynamically divided through the production area anomaly management unit (such as level 1 high-risk alarm, level 2 potential risk), reducing the interference of invalid alarms, achieving high timeliness, multi-source data fusion, intelligent hierarchical judgment, automated closed-loop management, resource service reuse and cost savings, and improving the adaptability to complex environments.

[0066] Embodiment 2

[0067] Based on Embodiment 1, as Figures 1 to 11 shown, an abnormal behavior analysis system for a power plant production area of the present invention further includes a drone inspection unit. The drone inspection unit includes a drone main body 101, propellers 102, support feet 103, and a housing 104. Multiple groups of propellers 102 are arranged on the drone main body 101. Two groups of support feet 103 are installed at the bottom end of the drone main body 101. A housing 104 is arranged at the bottom end of the drone main body 101. A square opening is arranged at the bottom end of the housing 104. The drone inspection unit further includes a sealing component, a detection component, a transmission component, and a gas analysis component. A circular hole is installed at each of the left and right ends of the housing 104. A sealing component is arranged on the housing 104 to seal the square opening. A detection component is installed inside the housing 104. A transmission component is arranged between the detection component and the sealing component. Two groups of gas analysis components are installed inside the housing 104;

[0068] The sealing component includes support plates 201, optical bars 202, springs 203, sealing plates 204, insertion rods 205, and square frames 206. Two pairs of support plates 201 are arranged at the bottom end of the housing 104. An optical bar 202 is arranged between each group of support plates 201. Two sealing plates 204 are respectively slidably connected to the two optical bars 202. Two springs 203 are respectively sleeved outside each optical bar 202. The two sealing plates 204 are both between the two springs 203. An insertion rod 205 is arranged at each of the two side ends of one sealing plate 204. A square frame 206 is installed at each of the two side ends of the other sealing plate 204. The insertion rod 205 is adapted to the square frame 206;

[0069] The detection component includes a servo electric cylinder 301, a lifting plate 302, an L-shaped guide plate 303, a smooth rod 304, a high-definition camera 305 and an infrared imager 306. At the top end inside the housing 104, a servo electric cylinder 301 is installed. At the bottom mobile end of the servo electric cylinder 301, a lifting plate 302 is installed. At the top end of the lifting plate 302, two groups of L-shaped guide plates 303 are provided. Two groups of smooth rods 304 are fixedly installed inside the housing 104. One group of L-shaped guide plates 303 is respectively slidably connected to one group of smooth rods 304. At the bottom end of the lifting plate 302, a high-definition camera 305 and an infrared imager 306 are respectively provided;

[0070] The transmission component includes a first bracket 401, a first fixed pulley 402, a second bracket 403, a second fixed pulley 404 and a rope 405. At the bottom end of the housing 104, two groups of first brackets 401 are provided. On each group of first brackets 401, a group of first fixed pulleys 402 are respectively rotatably provided. At the top end inside the housing 104, two groups of second brackets 403 are provided. On each group of second brackets 403, a group of second fixed pulleys 404 are respectively rotatably provided. The outer ends of the two groups of sealing plates 204 are respectively connected to the first ends of a group of ropes 405. The second ends of the two groups of ropes 405 respectively bypass the bottom ends of a group of first fixed pulleys 402 and enter the inside of the housing 104, bypass the top ends of the second fixed pulleys 404 and are connected to the top end of the lifting plate 302;

[0071] The gas analysis component includes a gas analyzer main body 501, an air delivery hose 502, a cylinder body 503, a suction head 504, a filter screen 505, a moving plate 506 and a support rod 507. At the rear end inside the housing 104, two groups of gas analyzer main bodies 501 are installed. The gas analyzer main body 501 is internally provided with an air extraction pump. At the input end of each group of gas analyzer main bodies 501, a group of air delivery hoses 502 are respectively provided. The input end of each group of air delivery hoses 502 is communicated with the output end of a group of cylinder bodies 503. At the top end of each group of sealing plates 204, a group of moving plates 506 are respectively provided. On each group of moving plates 506, a group of support rods 507 are respectively installed. The other ends of each group of support rods 507 are respectively installed with a group of cylinder bodies 503. At the input end of the cylinder body 503, a suction head 504 is provided. At the input end of the suction head 504, a filter screen 505 is installed. Notch openings are respectively provided on the left and right sides of the square opening;

[0072] It further includes a fixing plate 601, a rotating shaft 602 and a cover plate 603. At the left and right ends of the housing 104, two groups of fixing plates 601 are respectively provided. Between the two groups of fixing plates 601, a rotating shaft 602 is rotatably provided. On each group of rotating shafts 602, a group of cover plates 603 are respectively provided. A torsion spring is provided on the rotating shaft 602, and the torsion spring makes the cover plate 603 seal the circular hole;

[0073] At the closer ends of the two groups of sealing plates 204, buffer pads are respectively provided.

[0074] In this embodiment, under normal conditions, the two sets of sealing plates 204 are mirror-closer under the action of the elastic force of the corresponding springs 203. The two sets of sealing plates 204 cooperate with each other to seal the opposite ports. The lifting plate 302 synchronously pulls the two sets of sealing plates 204 through the guiding cooperation of the second fixed pulley 404 and the first fixed pulley 402 by the two sets of ropes 405, so that the two sets of sealing plates 204 perform a mirror-away operation. The lifting plate 302 drives the high-definition camera 305 and the infrared imager 306 to extend below the housing 104 through the square port for inspection operations. The sealing plate 204 drives a set of cylinders 503 to perform a mirror-away operation through the support rod 507 respectively, so that the suction heads 504 on each set of cylinders 503 extend outside the housing 104 through the round holes. The gas analyzer main body 501 is started to inhale the external gas by the suction head 504 and analyze it through the gas analyzer main body 501. After use, the servo electric cylinder 301 shortens, so that the lifting plate 302 rises. While the high-definition camera 305 and the infrared imager 306 enter the housing 104, the two sets of sealing plates 204 perform a mirror-closer operation under the action of the elastic force of the spring 203, so that while the two sets of sealing plates 204 seal the opposite ports, the two sets of suction heads 504 re-enter the housing 104 through the round holes for safety protection, and the two sets of cover plates 603 re-seal the round holes.

[0075] Embodiment 3

[0076] A method for analyzing abnormal behaviors in the production area of a power plant according to the present invention includes the following steps:

[0077] Step 1: Real-time collect video streams, face recognition data, and device sensor data through the 5G network;

[0078] Step 2: Use multi-threaded parallel processing technology to classify and analyze the data: extract frames from the video stream, input the behavior recognition model, output the probabilities of behaviors such as smoking and not wearing a safety helmet, combine the coordinates of the electronic fence and the personnel positioning data, and trigger an out-of-bounds alarm; analyze the device sensor data (such as temperature and vibration) to determine whether there is a fault or leakage;

[0079] Step 3: Alarm at different levels according to the analysis results:

[0080] Level 1 alarm (high-risk behavior, such as unauthorized entry into the high-voltage area): Immediately trigger a broadcast alarm and notify the safety officer;

[0081] Level 2 alarm (potential risk, such as not wearing a safety helmet): Push it to the management platform and record it;

[0082] Step 4: Store all data and generate a statistical report, supporting retrospective analysis by time, area, and behavior type.

[0083] The main functions realized by the present invention are:

[0084] 1. High real-time performance: The 5G network ensures low-latency video data transmission and supports second-level response.

[0085] 2. Automated management: Reduces manual intervention and improves safety management efficiency.

[0086] 3. Integration of 5G network and devices: Achieves high-bandwidth and low-latency data interaction.

[0087] 4. Multi-dimensional abnormal behavior analysis model: Integrates multiple algorithms to adapt to complex scenarios.

[0088] 5. Linkage alarm mechanism: Supports hierarchical response and automated disposal.

[0089] 6. Full-scenario coverage: Supports comprehensive monitoring of personnel behavior, device status, and environmental safety.

[0090] 7. Lower through the lifting plate 302, extend the high-definition camera 305 and the infrared imager 306 below the housing 104. The lifting plate 302 pulls the two sealing plates 204 through two groups of ropes 405, so that the two sealing plates 204 release the sealing of each other's openings. At the same time, the two sealing plates 204 use the moving plate 506 and the support rod 507 to extend the two suction heads 504 through the round holes to the outside of the housing 104 for collecting external gas.

[0091] For the abnormal behavior analysis method and system in the power plant production area of the present invention, its installation method, connection method, or setting method are all common mechanical methods, and any implementation that can achieve its beneficial effects can be carried out; the drone body 101, propeller 102, servo electric cylinder 301, high-definition camera 305, infrared imager 306, and gas analyzer body 501 of the abnormal behavior analysis method and system in the power plant production area of the present invention are purchased on the market. Those skilled in the art only need to install and operate according to the attached user manual, without the need for creative labor from those skilled in the art.

[0092] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art in the technical field, without departing from the technical principle of the present invention, several improvements and modifications can still be made, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims

1. An abnormal behavior analysis system for the production area of a power plant, comprising a data acquisition unit, a 5G network management unit, a data management unit, a personnel management unit, a device management unit, a behavior analysis unit, a production area abnormal management unit, and an abnormal alarm linkage management unit; characterized in that The data acquisition unit reuses the existing digital cameras, face recognition access control devices, and face capture cameras in the plant area, and adds intelligent cameras supporting 5G to collect video, personnel location, and equipment status data in real time and send them. The 5G network management unit is used to build a 5G local area network for the power plant, support low-latency transmission of video streams and interconnection of multiple devices, realize high-speed transmission of video data and device interconnection, and ensure real-time and clear video images. The data management unit uses the 5G local area network to receive the data of video, personnel location, and equipment status sent by the data acquisition unit, and performs data storage and playback. The personnel management unit is used to manage the regular employees, outsourced employees, and visitors in the plant area. The device management unit, with the support of the 5G local area network, conducts multi-item management of the plant area devices, including networking, basic information, and location-based management. The behavior analysis unit deploys a multi-dimensional abnormal behavior analysis model, including multi-dimensional analysis units for out-of-bounds, smoking, not wearing a safety helmet, falling, safety, fire protection, and equipment failure. The production area abnormal management unit is used to receive the analysis data sent by the behavior analysis unit. When it finds that the analysis result is abnormal and there is an alarm, it sends an electrical signal to the abnormal alarm linkage management unit. The abnormal alarm linkage management unit is used to receive the electrical signal sent by the production area abnormal management unit, realize local broadcast linkage, and broadcast the abnormality of a person or a device.

2. The abnormal behavior analysis system for the power plant production area according to claim 1, wherein The out-of-bounds analysis unit, based on personnel management and video analysis, and combined with the electronic fence and personnel positioning data, determines in real time whether a person enters a high-risk area, and pushes the analysis result to the production area abnormal management. If there is an alarm, it activates the abnormal alarm linkage management. The smoking analysis unit conducts out-of-bounds analysis based on personnel management and video analysis, and pushes the analysis result to the production area abnormal management. If there is an alarm, it activates the abnormal alarm linkage management. The not wearing a safety helmet analysis unit conducts not wearing a safety helmet analysis based on personnel management and video analysis, and pushes the analysis result to the production area abnormal management unit. If there is an alarm, it activates the abnormal alarm linkage management. The safety analysis unit is used to analyze the falling, fighting, and climbing of personnel and equipment failures, and pushes the analysis result to the production area abnormal management unit. If there is an alarm, it activates the abnormal alarm linkage management. The fire protection analysis unit is used to analyze equipment leakage, fire, etc. in the power plant area, and pushes the analysis result to the production area abnormal management unit. If there is an alarm, it activates the abnormal alarm linkage management.

3. The abnormal behavior analysis system for a power plant production area according to claim 1, characterized in that, It also includes a drone inspection unit. The drone inspection unit includes a drone main body (101), propellers (102), support feet (103) and a housing (104). Multiple groups of propellers (102) are arranged on the drone main body (101). Two groups of support feet (103) are installed at the bottom end of the drone main body (101). A housing (104) is arranged at the bottom end of the drone main body (101). A square opening is arranged at the bottom end of the housing (104). It also includes a sealing component, a detection component, a transmission component and a gas analysis component. A circular hole is installed at each of the left and right ends of the housing (104). A sealing component is arranged on the housing (104). The sealing component seals the square opening. A detection component is installed inside the housing (104). A transmission component is arranged between the detection component and the sealing component. Two gas analysis components are installed inside the housing (104).

4. The abnormal behavior analysis system for a power plant production area according to claim 3, characterized in that, The sealing component includes support plates (201), optical bars (202), springs (203), sealing plates (204), insertion rods (205) and square frames (206). Two pairs of support plates (201) are arranged at the bottom end of the housing (104). An optical bar (202) is arranged between each group of support plates (201). Two sealing plates (204) are respectively slidably connected to the two optical bars (202). Two springs (203) are respectively sleeved on the outer sides of each optical bar (202). The two sealing plates (204) are both between the two springs (203). An insertion rod (205) is arranged at each of the two side ends of one sealing plate (204). A square frame (206) is installed at each of the two side ends of the other sealing plate (204). The insertion rod (205) is adapted to the square frame (206).

5. The abnormal behavior analysis system for a power plant production area according to claim 3, wherein, The detection component includes a servo electric cylinder (301), a lifting plate (302), an L-shaped guide plate (303), an optical rod (304), a high-definition camera (305) and an infrared imager (306). A servo electric cylinder (301) is installed at the top end inside the housing (104). The bottom mobile end of the servo electric cylinder (301) is installed with a lifting plate (302). Two L-shaped guide plates (303) are arranged at the top end of the lifting plate (302). Two optical rods (304) are fixedly installed inside the housing (104). One L-shaped guide plate (303) is respectively slidably connected to one optical rod (304). A high-definition camera (305) and an infrared imager (306) are respectively arranged at the bottom end of the lifting plate (302).

6. The abnormal behavior analysis system for a power plant production area according to claim 5, wherein, The transmission assembly includes a first bracket (401), a first fixed pulley (402), a second bracket (403), a second fixed pulley (404) and a rope (405). Two groups of first brackets (401) are provided at the bottom end of the housing (104), and a group of first fixed pulleys (402) are rotatably provided on each group of first brackets (401). Two groups of second brackets (403) are provided at the inner top end of the housing (104), and a group of second fixed pulleys (404) are rotatably provided on each group of second brackets (403). The outer ends of the two groups of sealing plates (204) are respectively connected to the first ends of a group of ropes (405). The second ends of the two groups of ropes (405) respectively bypass the bottom ends of a group of first fixed pulleys (402) and enter the interior of the housing (104), bypass the top ends of the second fixed pulleys (404) and are connected to the top end of the lifting plate (302).

7. The abnormal behavior analysis system for a power plant production area according to claim 4, wherein The gas analysis assembly includes a gas analyzer main body (501), an air delivery hose (502), a cylinder body (503), a suction head (504), a filter net (505), a moving plate (506) and a support rod (507). Two groups of gas analyzer main bodies (501) are installed at the rear end inside the housing (104). An air delivery hose (502) is provided at the input end of each group of gas analyzer main bodies (501). The input end of each air delivery hose (502) is communicated with the output end of a cylinder body (503). A moving plate (506) is provided at the top end of each group of sealing plates (204). A support rod (507) is installed on each group of moving plates (506). The other ends of each group of support rods (507) are respectively installed with a cylinder body (503). The input end of the cylinder body (503) is provided with a suction head (504). A filter net (505) is installed at the input end of the suction head (504). Notch openings are respectively provided on the left and right sides of the square opening.

8. The abnormal behavior analysis system for the power plant production area according to claim 3, wherein It further includes a fixing plate (601), a rotating shaft (602) and a cover plate (603). Two groups of fixing plates (601) are respectively provided at the left and right ends of the housing (104). A rotating shaft (602) is rotatably provided between the two groups of fixing plates (601). A cover plate (603) is provided on each group of rotating shafts (602). A torsion spring is provided on the rotating shaft (602), and the torsion spring seals the round hole by the cover plate (603).

9. The abnormal behavior analysis system for a power plant production area according to claim 4, wherein, Buffer pads are respectively provided at the ends of the two groups of sealing plates (204) close to each other.

10. A method for analyzing abnormal behaviors in the production area of a power plant, characterized in that, It includes the following steps: Step 1: Real-time collect video streams, face recognition data and device sensor data through the 5G network; Step 2: Use multi-thread parallel processing technology to classify and analyze the data: extract frames from the video stream, input the behavior recognition model, output the behavior probabilities of smoking and not wearing a safety helmet, combine the electronic fence coordinates with the personnel positioning data to trigger out-of-bounds alarms; analyze the device sensor data to determine whether there is a fault or leakage; Step 3: Alarm at different levels according to the analysis results: Level 1 alarm: Immediately trigger a broadcast alarm and notify the safety officer; Level 2 alarm: Push it to the management platform and record it; Step 4: Store all data and generate a statistical report, supporting retrospective analysis by time, area and behavior type.

Images