A Smart Alarm and Control Method for Personnel Safety in Power Plants

Abstract

This invention discloses an intelligent alarm and control method for personnel safety in power plants, relating to the field of power industry safety control technology. The method includes: acquiring multimodal sensing data related to personnel based on the existing sensing system of the power plant; using a multimodal analysis model to fuse and analyze the collected multimodal sensing data to identify personnel safety risks; generating an alarm trigger signal when the risk level reaches a preset threshold; combining intelligent agent technology to generate a personalized alarm strategy after receiving the alarm trigger signal, simultaneously finding the nearest alarm device based on the real-time location of the personnel for continuous alarm, and establishing a continuous monitoring mechanism to track personnel status in real time; continuously monitoring the behavior and location changes of the alarmed personnel until the risk is confirmed to be completely eliminated. This invention enables cross-system correlation analysis, improves the accuracy of risk identification, effectively shortens response time, and improves the safety level of power plants.

Description

Technical Field

[0001] This invention relates to the field of safety management and control technology in the power industry, and in particular to an intelligent alarm and control method for personnel safety in power plants. Background Technology

[0002] With the rapid development of the power industry, power plants are expanding in scale and the production environment is becoming increasingly complex, posing serious challenges to personnel safety management. Traditional personnel safety management methods mainly rely on manual inspections and video surveillance playback, which have significant drawbacks such as poor timeliness, weak systemicity, low level of intelligence, and lack of closed-loop management.

[0003] In existing technologies, power plants typically deploy multiple independent information systems, such as access control systems, security systems, personnel positioning systems, monitoring systems, ticketing systems, and maintenance systems. However, data silos exist between these systems, lacking effective information linkage mechanisms and hindering comprehensive risk assessment across systems. Furthermore, existing video surveillance systems rely heavily on manual monitoring, resulting in high workload, low efficiency, and susceptibility to missed detections due to fatigue. Traditional image recognition algorithms have limited accuracy in recognizing complex scenes and personnel in various poses, making them ill-suited to the complex and ever-changing operating environment of power plants. In recent years, with the rapid development of artificial intelligence, technologies such as multimodal large models, intelligent agent technology, and knowledge graphs have matured, providing new technical pathways to address these issues. However, how to deeply integrate these advanced technologies with the actual business scenarios of power plants to construct an intelligent personnel safety management system adapted to the characteristics of the power industry remains a pressing technical challenge. Summary of the Invention

[0004] In view of the problems existing in current intelligent alarm and control methods for personnel safety in power plants, this invention is proposed. This invention aims to overcome the shortcomings of the prior art and provide an intelligent alarm and control method for personnel safety in power plants. By constructing a multimodal perception fusion system, an intelligent analysis and decision-making system, and a closed-loop control and execution system, it achieves automatic identification, intelligent judgment, proactive alarm, and continuous tracking of unsafe behaviors of personnel in power plants until the risk is completely eliminated, thereby significantly improving the timeliness, accuracy, and reliability of personnel safety control in power plants. Therefore, the problem this invention aims to solve is how to provide an intelligent alarm and control method for personnel safety in power plants.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0006] In a first aspect, the present invention provides a method for intelligent alarm and control of personnel safety in power plants, comprising: acquiring multimodal sensing data related to personnel based on the existing sensing system of the power plant; the sensing system includes an access control system, a security system, a personnel information system, a personnel positioning system, a monitoring system, a ticketing system, a maintenance system, and an inspection system;

[0007] The multimodal analysis model is used to fuse and analyze the collected multimodal perception data to identify personnel safety risks. Combined with the safety management knowledge base and large language model, the personnel safety risks are comprehensively judged, and an alarm trigger signal is generated when the risk level reaches a preset threshold.

[0008] After receiving an alarm trigger signal, the intelligent agent technology integrates basic personnel information, real-time location data, related work order information, behavior analysis results, and historical alarm records to generate personalized alarm strategies, send alarm reminders to managers, and continuously alarm the nearest alarm device based on the real-time location of the personnel, and establishes a continuous monitoring mechanism to track the personnel status in real time.

[0009] Continuously monitor the behavior and location changes of the accused personnel until the risk is confirmed to be completely eliminated, stop the alarm and generate a complete handling record for intelligent personnel safety management.

[0010] As a preferred embodiment of the intelligent alarm and control method for personnel safety in power plants described in this invention, the multimodal sensing data includes text modality, audio modality, video modality, location modality, and time-series modality; the text modality includes basic personnel information, work order information, and maintenance records; the audio modality includes ambient sound; the video modality includes real-time video streams from monitoring cameras; the location modality includes UWB / BeiDou / GPS positioning coordinates and electronic fence area codes; and the time-series modality includes personnel trajectory data.

[0011] As a preferred embodiment of the intelligent alarm and control method for personnel safety in power plants described in this invention, the multimodal analysis model includes a collaborative architecture of small model + large model;

[0012] The small model layer deploys dedicated detection models, including object detection models, image classification models, people tracking algorithms, and face recognition models; the large model layer introduces a multimodal large language model to perform semantic understanding and comprehensive reasoning on the structured features, text work order information, and knowledge base rules output by the small model layer, and to conduct semantic-level risk assessment of the scenario.

[0013] As a preferred embodiment of the intelligent alarm and control method for personnel safety in power plants described in this invention, the personalized alarm generation strategy includes:

[0014] The personnel who triggered the alarm were uniquely identified and their identity attributes, job categories, work qualifications, historical work records, and safety training records were extracted from the personnel information system.

[0015] Real-time spatial coordinates are obtained from the personnel positioning system and mapped to the work area, equipment unit and risk area identifiers by combining the spatial topology model of the power plant.

[0016] Extract work order numbers, job types, job time windows, permit status, and safety measure implementation status from the two-ticket system and the maintenance system; extract recent inspection routes and inspection results from the inspection system.

[0017] The behavioral analysis results of the monitoring and security systems are analyzed, and abnormal behavior tags and violation operation records are transformed into standardized behavioral feature vectors. These vectors are then matched and analyzed with work order requirements to calculate behavioral deviation indicators.

[0018] As a preferred embodiment of the intelligent alarm and control method for personnel safety in power plants described in this invention, the calculation formula for the behavioral deviation index is:

[0019]

[0020] in, Indicators representing behavioral deviations Indicates the first Actual behavioral characteristic value, Indicates the first The standard operating procedure characteristic value of the actual behavior. Weighting coefficients representing behavioral characteristics This represents the total number of behavioral characteristics.

[0021] As a preferred embodiment of the intelligent alarm and control method for personnel safety in power plants according to the present invention, the step of finding the nearest alarm device based on the real-time location of personnel includes: calculating the spatial distance from the current location of personnel to each alarm device based on the GIS map of the power plant and the alarm device distribution data, combined with the real-time location coordinates of personnel, as expressed as:

[0022]

[0023] in, Indicates the personnel's position and the number Spatial distance between alarm devices Indicates the current coordinates of the personnel. Indicates the first The coordinates of each alarm device;

[0024] After calculating the distance to all devices, the device with the smallest distance is selected as the target alarm device. Taking into account device type, device status and environmental obstruction factors, the optimal combination of alarm devices is selected for continuous alarm.

[0025] As a preferred embodiment of the intelligent alarm and control method for personnel safety in power plants according to the present invention, the intelligent control of personnel safety includes:

[0026] Establish a continuous monitoring mechanism to update personnel location and behavior data at fixed time intervals, and recalculate behavioral deviation indicators and risk status.

[0027] A multi-condition joint judgment mechanism is adopted to confirm risk elimination, and the multi-conditions include location conditions, behavioral conditions and system conditions.

[0028] The location condition is that the personnel's location coordinates have left the danger zone defined by the electronic fence; the behavior condition is that video analysis confirms that the personnel have corrected their unsafe behavior; the system condition is that the relevant work order status has been updated or the access control system records that the personnel have left the specific area; when all conditions are met, the intelligent agent determines that the risk has been eliminated, automatically stops the alarm, and generates a complete handling record that includes the alarm time, location, risk type, handling process, and elimination time.

[0029] Secondly, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of a method for intelligent alarm and control of personnel safety in power plants.

[0030] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements the steps of a method for intelligent alarm and control of personnel safety in power plants.

[0031] The beneficial effects of this invention are as follows: This invention integrates multi-source data such as access control, security, positioning, and operation systems to construct a unified personnel safety perception, realize cross-system correlation analysis, and improve the accuracy of risk identification; through the collaboration of real-time detection by small models and deep inference by large models, it balances response efficiency and the ability to understand complex scenarios, reducing false alarms and missed alarms; by introducing intelligent agents to realize autonomous decision-making and dynamic alarms, and by combining location information to optimize alarm methods and continuously track, a closed-loop management mechanism is formed, transforming traditional post-event handling into pre-event warning and in-event intervention, effectively shortening response time and improving the safety level of power generation stations. Attached Figure Description

[0032] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 This is a flowchart of an intelligent alarm and control method for personnel safety in power plants. Detailed Implementation

[0034] To make the above-mentioned objects, features, and advantages of the present invention more readily understood, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0035] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0036] Secondly, the term "one embodiment" or "example" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the invention. An embodiment appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment that selectively excludes other embodiments.

[0037] Example 1

[0038] Reference Figure 1 This is the first embodiment of the present invention, which provides a method for intelligent alarm and control of personnel safety in power plants, including:

[0039] S1. Based on the existing sensing system of the power station, access multimodal sensing data related to personnel; the sensing system is designed for the characteristics of power stations and integrates multiple heterogeneous systems such as access control system, security system, personnel information system, personnel positioning system, monitoring system, ticketing system, maintenance system, and inspection system, and collects multi-source multimodal sensing information related to personnel in each system. The information modalities include text, audio, video, location coordinates, and time-series modalities.

[0040] S2. Utilize a multimodal analysis model to fuse and analyze the collected perception data to identify personnel safety risks. Based on the information collected by the personnel perception system and relying on a professional analysis model, the multimodal analysis system realizes functions such as personnel location tracking, unsafe behavior identification, work area misalignment judgment, personnel identity verification, cross-camera tracking and re-identification. Combined with a safety management knowledge base and a large language model, it comprehensively assesses the safety risks that may be caused by personnel's own behavior or the work environment, and generates an alarm trigger signal when the risk level reaches a preset threshold.

[0041] S3. Relying on intelligent agent technology, when an alarm trigger signal is received, the alarm and control process is automatically executed. The intelligent agent integrates basic personnel information, real-time location data, related work order information, behavior analysis results and historical alarm records, and combines multimodal large model, security control rule base and site geographic information database components to generate personalized alarm strategies.

[0042] The intelligent agent automatically sends alarm reminders to managers, retrieves relevant perception data of personnel as evidence as needed, and generates standardized alarm information by combining it with the safety management knowledge base. Based on the real-time location information of personnel, it actively calculates and finds the alarm device closest to the personnel, issues continuous alarms through various means such as sound and light alarms, intelligent broadcasts, and mobile terminals, and establishes a continuous monitoring mechanism to track the status of personnel in real time.

[0043] S4. Continuously monitor the behavior and location changes of the personnel being alerted until the risk is confirmed to be completely eliminated. When it is detected that the personnel have left the dangerous area, are wearing necessary protective equipment, or have corrected unsafe behavior, the intelligent agent determines that the risk has been eliminated, automatically stops the alarm, and generates a complete alarm handling record, including alarm time, location, risk type, handling process, elimination time, and other elements, to achieve closed-loop management of the entire process of discovery-alarm-de-confirmation.

[0044] Specifically, multi-source, multi-modal sensing information includes: text modality (basic personnel information, work order information, maintenance records), video modality (real-time video stream from surveillance cameras), audio modality (on-site environmental sound), location modality (UWB / BeiDou / GPS positioning coordinates, electronic fence area coding), and time-series modality (personnel trajectory data), etc.

[0045] The multimodal analysis model adopts a collaborative architecture of small model + large model: the small model layer deploys lightweight dedicated detection models, including object detection models, image classification models, people tracking algorithms, face recognition models, etc., to achieve millisecond-level real-time analysis;

[0046] The large model layer introduces a multimodal large language model to perform semantic understanding and comprehensive reasoning on the structured features, text work order information, and knowledge base rules output by the small model, thereby achieving semantic-level risk assessment in complex scenarios. The two collaborate asynchronously through message queues, improving analysis accuracy while ensuring real-time performance.

[0047] The intelligent agent possesses perception, memory, reasoning, execution, and tool usage capabilities. It can call upon the geographic information system to calculate the optimal alarm equipment, call upon the knowledge base to query safety procedures, and call upon the work order system to verify work permits, etc.

[0048] Finding the nearest alarm devices specifically involves: the intelligent agent calculating the distance from the current location of the personnel to each alarm device based on the site's GIS map and alarm device deployment data, combined with the personnel's real-time location coordinates, and comprehensively considering factors such as device type, device status, and environmental obstruction, to select the optimal combination of alarm devices and achieve efficient delivery of alarm information.

[0049] Upon receiving an alarm trigger signal, the system first performs unique identifier resolution on the personnel who triggered the signal, and then extracts basic information data of the personnel from the personnel information system based on this identifier, including identity attributes, job category, work qualifications, historical work records, and safety training records. At the same time, the system obtains the real-time spatial coordinates of the personnel from the personnel positioning system, and combines them with the spatial topology model of the power plant to map the personnel's positioning information to specific work areas, equipment units, and risk area identifiers, thereby forming a personnel location description with spatial semantics.

[0050] Based on the acquisition of basic personnel information and real-time location data, work order information related to the current work status of personnel is extracted from the two-ticket system and the maintenance system. This includes the currently executed operation ticket or work ticket number, work type, work time window, permission status, and safety measure implementation status. The work order information is then structured to form a unified data representation. At the same time, the recent inspection routes and inspection results of personnel are extracted from the inspection system.

[0051] The behavioral analysis results collected by the monitoring and security systems are analyzed, and the video recognition results, abnormal behavior tags, and violation operation records are transformed into standardized behavioral feature vectors. These vectors are then matched with the current work order requirements of the personnel to quantify the degree of deviation between the personnel's actual behavior and the standardized work behavior, forming a behavioral deviation index, expressed as follows:

[0052]

[0053] in, Indicators representing behavioral deviations Indicates the first Actual behavioral characteristic value, Indicates the first The standard operating procedure characteristic value of the actual behavior. Weighting coefficients representing behavioral characteristics Indicates the total number of behavioral characteristics;

[0054] After completing the construction of behavioral characteristics, basic personnel information, real-time location data, work order information, and behavioral deviation indicators are integrated to form a multi-dimensional personnel status description vector. Combined with historical alarm records, the risk evolution trend of personnel is analyzed. Historical alarm records include historical alarm types, occurrence frequency, processing results, and response time. By performing time series modeling on historical data, personnel risk accumulation factors are obtained.

[0055] Based on intelligent agent technology, the system receives and integrates personnel status data, combines it with a pre-built set of security control rules, performs policy matching on current risks, and generates personalized alarm policies for personnel.

[0056] The strategy includes alarm level, alarm method, alarm frequency, and response requirements. Based on the current spatial location of personnel, it uses a spatial distance calculation method to select the nearest alarm device from the power plant's alarm device set, as shown below:

[0057]

[0058] in, Indicates the personnel's position and the number Spatial distance between alarm devices Indicates the current coordinates of the personnel. Indicates the first The coordinates of each alarm device are calculated, and the device with the minimum distance is selected as the target alarm device.

[0059] After identifying the alarm device, an alarm reminder message is sent to the management terminal. The alarm content includes the personnel's identity, risk type, current location, behavioral deviation, and suggested handling measures. At the same time, an alarm command is issued to the target alarm device to trigger an alarm according to the set alarm strategy.

[0060] After an alarm is triggered, a continuous monitoring mechanism is established to update the personnel's location and behavior data at fixed time intervals, and to recalculate the behavior deviation indicators and risk status, so as to realize dynamic tracking of personnel status and real-time adjustment of alarm strategies.

[0061] Risk elimination confirmation adopts a multi-condition joint judgment mechanism: location conditions (personnel location coordinates leave the danger zone defined by the electronic fence), behavioral conditions (video analysis confirms that the personnel have corrected unsafe behaviors), system conditions (relevant work order status updates, access control system records that the personnel have left the specific area), etc.; the intelligent agent integrates the above conditions and determines the risk status through rule engine or model reasoning to avoid false alarms and missed alarms.

[0062] This embodiment also provides a computer device applicable to a method for intelligent alarm and control of personnel safety in power plants, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement all or part of the steps of the method described in the above embodiments of the present invention.

[0063] This embodiment also provides a storage medium storing a computer program thereon. When the computer program is executed by a processor, it performs the method in any optional implementation of the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0064] The storage medium proposed in this embodiment and the data storage method proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0065] Example 2

[0066] This embodiment uses a wind farm booster station as an example to explain in detail the closed-loop control process of the present invention.

[0067] Scenario setting: The 35kV distribution room of a wind farm's booster station is a live-lined area with an electronic fence, allowing only personnel with valid work permits and high-voltage electrician certificates to enter during permitted time periods.

[0068] S1, 9:15:32, Personnel Li Si entered the substation by swiping his access card; 9:16:45, the UWB positioning system detected that Li Si had entered the electronic fence area of ​​the 35kV power distribution room; the area camera captured the video stream; the system queried the two-ticket system in real time and found that there was no valid work ticket for the 35kV power distribution room at the current time.

[0069] S2. The video analysis module detected a person target but did not detect the wearing of a safety helmet. The location analysis module confirmed that Li Si was in the live area. The text analysis module confirmed that Li Si did not have a valid work permit. The fusion reasoning module output a high-risk alarm.

[0070] S3. The intelligent agent receives high-risk signals, loads context information, queries the knowledge base to generate alarm strategies, calls GIS services to find the most recent alarm devices, activates the audible and visual alarms and intelligent broadcasts to issue warnings to drive away the devices, pushes alarm notifications to the duty manager's APP, and starts continuous monitoring.

[0071] S4, 9:17:20, the positioning system shows that Li Si has started moving towards the exit; 9:17:45, video analysis confirms that Li Si has put on a safety helmet and left the switch cabinet; 9:18:10, the positioning system shows that Li Si has left the electronic fence area; 9:18:30, the access control system records that Li Si left the booster station; the intelligent agent comprehensively determines that the risk has been eliminated, stops the alarm and generates a complete alarm record.

[0072] In summary, this invention breaks down traditional information silos, deeply integrating data from multiple heterogeneous systems such as access control, security, positioning, monitoring, ticketing, and maintenance to construct a unified view of personnel safety perception. This enables cross-system correlation analysis and comprehensive judgment, significantly improving the comprehensiveness and accuracy of risk identification. This invention employs a collaborative architecture of real-time detection using a small model and deep inference using a large model. The small model ensures millisecond-level response speed, meeting real-time requirements; the large model provides semantic-level understanding capabilities to handle complex scenarios. Together, they achieve 24 / 7 automated monitoring, completely eliminating the limitations of manual monitoring. This invention fully leverages the efficiency of the small model in specific tasks and the advantages of the large model in general reasoning. Through a layered architecture design, it ensures the efficiency of real-time detection while effectively reducing false alarm and false negative rates and improving identification accuracy in complex scenarios through the knowledge injection and logical reasoning capabilities of the large model. This invention introduces intelligent agent technology, endowing the system with autonomous decision-making and proactive intervention capabilities. It can not only automatically detect risks but also proactively select the optimal alarm method based on personnel location, continuously tracking until the risk is eliminated, forming a complete closed-loop management record, and realizing a shift from passive monitoring to proactive prevention. This invention transforms traditional post-event tracing into pre-event early warning and in-event intervention through multimodal fusion perception, intelligent agent autonomous decision-making, and closed-loop control execution, significantly shortening risk response time, reducing the probability of safety accidents, and improving the intrinsic safety level of power generation stations.

[0073] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for intelligent alarm and control of personnel safety in power plants, characterized in that: include, Based on the existing sensing system of the power plant, multimodal sensing data related to personnel is acquired; the sensing system includes an access control system, a security system, a personnel information system, a personnel positioning system, a monitoring system, a ticketing system, a maintenance system, and an inspection system; The multimodal analysis model is used to fuse and analyze the collected multimodal perception data to identify personnel safety risks. Combined with the safety management knowledge base and large language model, the personnel safety risks are comprehensively judged, and an alarm trigger signal is generated when the risk level reaches a preset threshold. After receiving an alarm trigger signal, the intelligent agent technology integrates basic personnel information, real-time location data, related work order information, behavior analysis results, and historical alarm records to generate personalized alarm strategies, send alarm reminders to managers, and continuously alarm the nearest alarm device based on the real-time location of the personnel, and establishes a continuous monitoring mechanism to track the personnel status in real time. Continuously monitor the behavior and location changes of the accused personnel until the risk is confirmed to be completely eliminated, stop the alarm and generate a complete handling record for intelligent personnel safety management.

2. The intelligent alarm and control method for personnel safety in power plants as described in claim 1, characterized in that: The multimodal sensing data includes text modality, audio modality, video modality, location modality, and time-series modality; text modality includes basic personnel information, work order information, and maintenance records; audio modality includes ambient sound; video modality includes real-time video streams from surveillance cameras; location modality includes UWB / BeiDou / GPS positioning coordinates and electronic fence area codes; and time-series modality includes personnel trajectory data.

3. The intelligent alarm and control method for personnel safety in power plants as described in claim 1, characterized in that: The multimodal analysis model includes a collaborative architecture of small model + large model; The small model layer deploys dedicated detection models, including object detection models, image classification models, people tracking algorithms, and face recognition models; the large model layer introduces a multimodal large language model to perform semantic understanding and comprehensive reasoning on the structured features, text work order information, and knowledge base rules output by the small model layer, and to conduct semantic-level risk assessment of the scenario.

4. The intelligent alarm and control method for personnel safety in power plants as described in claim 1, characterized in that: The personalized alert generation strategy includes: The personnel who triggered the alarm were uniquely identified and their identity attributes, job categories, work qualifications, historical work records, and safety training records were extracted from the personnel information system. Real-time spatial coordinates are obtained from the personnel positioning system and mapped to the work area, equipment unit and risk area identifiers by combining the spatial topology model of the power plant. Extract work order numbers, job types, job time windows, permit status, and safety measure implementation status from the two-ticket system and the maintenance system; extract recent inspection routes and inspection results from the inspection system. The behavioral analysis results of the monitoring and security systems are analyzed, and abnormal behavior tags and violation operation records are transformed into standardized behavioral feature vectors. These vectors are then matched and analyzed with work order requirements to calculate behavioral deviation indicators.

5. The intelligent alarm and control method for personnel safety in power plants as described in claim 4, characterized in that: The formula for calculating the behavioral deviation index is as follows: ; in, Indicators representing behavioral deviations Indicates the first Actual behavioral characteristic value, Indicates the first The standard operating procedure characteristic value of the actual behavior. Weighting coefficients representing behavioral characteristics This represents the total number of behavioral characteristics.

6. The intelligent alarm and control method for personnel safety in power plants as described in claim 1, characterized in that: The method of locating the nearest alarm device based on the real-time location of personnel includes: calculating the spatial distance from the current location of a person to each alarm device based on the GIS map of the power plant and the alarm device deployment data, combined with the real-time location coordinates of the personnel, and expressing it as follows: ; in, Indicates the personnel's position and the number Spatial distance between alarm devices Indicates the current coordinates of the personnel. Indicates the first The coordinates of each alarm device; After calculating the distance to all devices, the device with the smallest distance is selected as the target alarm device. Taking into account device type, device status and environmental obstruction factors, the optimal combination of alarm devices is selected for continuous alarm.

7. The intelligent alarm and control method for personnel safety in power plants as described in claim 1, characterized in that: The aforementioned intelligent management and control of personnel safety includes: Establish a continuous monitoring mechanism to update personnel location and behavior data at fixed time intervals, and recalculate behavioral deviation indicators and risk status; A multi-condition joint judgment mechanism is adopted to confirm risk elimination, and the multi-conditions include location conditions, behavioral conditions and system conditions. The location condition is that the personnel's location coordinates have left the danger zone defined by the electronic fence; the behavior condition is that video analysis confirms that the personnel have corrected their unsafe behavior; the system condition is that the relevant work order status has been updated or the access control system records that the personnel have left the specific area; when all conditions are met, the intelligent agent determines that the risk has been eliminated, automatically stops the alarm, and generates a complete handling record that includes the alarm time, location, risk type, handling process, and elimination time.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the intelligent alarm and control method for personnel safety in power plants as described in any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the intelligent alarm and control method for personnel safety in power plants as described in any one of claims 1 to 7.

Images