Intelligent dynamic safety management and control system and method for electric power outsourcing operation

Abstract

The invention relates to the technical field of operation safety monitoring, and discloses an intelligent dynamic safety management and control system and method for electric power outsourcing operation, and the system comprises an operation analysis module which is used for decomposing obtained operation information into current operation elements, and generating a dynamic risk thermodynamic diagram of the current operation; the operation elements comprise operation steps, operation space, operation resources and operation risks; the perception scheduling module is used for scheduling and configuring a monitoring area and perception resources in real time according to the dynamic risk thermodynamic diagram, and constructing a current management and control area; acquiring state information of the operating personnel in real time based on the generated management and control area; and the risk prediction module is used for obtaining a prediction risk coefficient based on the coupling risk coefficient and the operation intention on the basis of the coupling risk coefficient based on the operation elements, the analysis state information and the coupling risk coefficient of the operation stage, and outputting intervention measures according to risk prediction. According to the method, the problem of safety control failure caused by large monitoring blind area, weak identification capability and response delay in electric power outsourcing operation is effectively solved.

Description

Technical Field

[0001] This invention relates to the field of operational safety monitoring technology, specifically to an intelligent dynamic safety management and control system and method for power outsourcing operations. Background Technology

[0002] In the power generation sector, outsourcing has become a crucial link in ensuring equipment operation, maintenance, and construction. However, this type of work, characterized by high personnel turnover, varying safety skills, and complex and ever-changing work scenarios, poses systemic risks to safe production. To address this challenge, the industry generally relies on technical safeguards such as video surveillance systems, personnel positioning systems, and environmental sensors. However, existing technologies are fragmented and static in practical applications, revealing serious shortcomings when dealing with dynamic and high-risk outsourcing operations.

[0003] First, existing factory safety monitoring mainly relies on pre-installed fixed cameras, whose field of view and location are difficult to change flexibly. Outsourced work, such as wiring at heights, confined space maintenance, and pipeline corridor inspections, is highly mobile and temporary, with work sites often located outside the field of view of fixed cameras. Statistics show that over 60% of high-risk work areas in the factory have blind spots, leaving many high-risk operations invisible and uncontrollable, constituting a fundamental loophole in safety management.

[0004] Secondly, traditional video analysis has an accuracy rate of less than 50% in identifying key behaviors such as not wearing safety ropes and improper use of tools; while conventional positioning technology has an error of up to ±3 meters in complex steel frame environments, making it impossible to accurately associate personnel locations with specific electrified intervals or edge openings, resulting in inaccurate risk warnings and ineffective control.

[0005] Furthermore, when the monitoring center detects potential risks, it typically issues voice alerts via the broadcast system. However, in high-noise working environments, over 80% of these voice alerts are drowned out, failing to effectively reach the target personnel. Moreover, the process from risk detection to on-site handling requires multiple levels of confirmation, with an average response time exceeding 8 minutes, missing the critical window for intervention.

[0006] In summary, existing technologies are no longer adequate to meet the safety management needs of outsourced operations, which are characterized by changing scenarios and dynamic risks. This results in passive and inefficient on-site supervision and weak accident prevention capabilities. Summary of the Invention

[0007] The present invention aims to provide an intelligent dynamic safety management and control system and method for power outsourcing operations, in order to solve the problems of low coverage, inaccurate identification, and slow response of existing monitoring technologies.

[0008] To achieve the above objectives, the present invention employs the following technical solution for an intelligent dynamic safety management and control system for power outsourcing operations, comprising: The task parsing module is used to decompose the acquired task information into current task elements and generate a dynamic risk heat map of the current task; the task elements include task steps, task space, task resources and task risks; The perception and scheduling module is used to schedule and configure monitoring areas and perception resources in real time based on dynamic risk heat maps, and construct the current control zone; and to obtain the status information of operators in real time based on the generated control zone; The risk prediction module is used to analyze the coupling risk coefficient between status information and work stage based on work elements, obtain the predicted risk coefficient based on the coupling risk coefficient and work intention, and output intervention measures based on the risk prediction.

[0009] Meanwhile, this solution also provides an intelligent dynamic safety management and control method for power outsourcing operations, applied to the aforementioned intelligent dynamic safety management and control system for power outsourcing operations, including the following steps: S1: Obtain current task information and decompose the task information into four-tuple task elements; generate dynamic risk heatmaps for each task step based on the task elements; S2, based on a dynamic risk heat map, combines the current work steps with the scheduling and configuration of monitoring areas and sensing resources to construct the current control zone; and obtains the status information of the workers in real time; S3 calculates the coupling risk coefficient in real time based on the work elements, current steps, and personnel status information, and calculates the predicted risk coefficient in combination with the work intention to perform risk prediction. S4, based on risk prediction, output operational intervention measures according to intervention rules.

[0010] The principles and advantages of this scheme are: This solution breaks away from the traditional framework of static monitoring, no longer viewing safety control as passive surveillance of moving targets within a fixed space, but rather as intelligent tracking and judgment of the dynamic work process itself. By analyzing work orders, the system decomposes abstract work tasks into structured steps, spaces, resources, and risk elements, generating a risk heat map that dynamically changes and shifts as the work progresses. This map is not a pre-drawn topographic map, but a "risk weather cloud map" rendered in real time along with the work rhythm, accurately reflecting "where, when, and due to what work activity, what kind of risk exists."

[0011] Based on this dynamic risk map, the system flexibly schedules mobile and fixed sensing resources. These resources are not evenly distributed but effectively allocated according to risk levels, forming adaptive control zones to ensure the most critical links are under the most stringent monitoring. More importantly, by real-time coupling analysis of personnel behavior and the standardized requirements of current work procedures, the system not only identifies immediate deviations but also predicts future risk trends based on work intentions. This achieves a leap from "post-event alarms" to "in-process intervention" and then to "pre-event prevention." Its intervention measures are therefore no longer general audible and visual alarms but precise guidance tied to specific risk types, work procedures, and even personnel roles, such as prompting tool verification, warning of incorrect step sequences, or forcibly preventing accidental entry into live areas.

[0012] This solution systematically addresses the long-standing problems in power outsourcing operations, such as large blind spots in static monitoring, shallow risk identification, and slow alarm response. It increases safety control coverage from less than 40% to near-complete coverage, improves the accuracy of risk behavior identification and location to the centimeter and step level, transforms ineffective broadcast alarms into effective interventions directly reaching individuals, and reduces the average emergency response time from over 8 minutes to minutes or even seconds. Ultimately, this solution transforms safety management from a lagging, fragmented supervisory task into a proactive assurance capability embedded in the production process, achieving inherently safe control over the risks of dynamic and complex outsourced operations, significantly reducing the accident rate, and improving overall operational efficiency. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of the intelligent dynamic safety management and control system for power outsourcing operations according to the present invention; Figure 2 This is a flowchart illustrating the intelligent dynamic safety management and control method for power outsourcing operations according to the present invention. Detailed Implementation

[0014] The following detailed description illustrates the specific implementation method: The intelligent dynamic safety management and control system and method for power outsourcing operations in this embodiment transforms safety management from monitoring static scenarios to intelligent monitoring of work processes by constructing a dynamic coupling model of "operation-risk". This effectively solves the problems of large blind spots, shallow identification and slow response in traditional monitoring.

[0015] Option 1 An intelligent dynamic safety management and control system for power outsourcing operations is provided, as shown in the attached document. Figure 1 As shown, it includes: The task parsing module is used to decompose the acquired task information into current task elements and generate a dynamic risk heat map of the current task; the task elements include task steps, task space, task resources and task risks.

[0016] In this embodiment, key information is extracted from the electronic work order. A pre-trained model (such as BERT) can be used for Named Entity Recognition (NER) to extract work steps, required tools, personnel requirements, risk descriptions, etc. This information is then mapped to four work elements, transforming unstructured work information, such as work orders and operation tickets, into structured work elements, forming an element set E, represented as follows: ;in, The work steps are represented as follows Each step Includes description and standard operating time. and the set of precursor steps; The workspace is represented as Each step is associated with a three-dimensional spatial region, then For the permission threshold of this step, such as ; A point in space; It is a three-dimensional spatial region; , This defines a polyhedral region.

[0017] Job resources are represented as In the formula, Personnel qualification matrix; A tool list vector; This is a vector for the list of protective equipment.

[0018] Operational risks are represented as Each .

[0019] In this embodiment, the spatial distribution of risk is not fixed, but rather varies with the current operational steps. and its spatial threshold This is jointly determined; therefore, based on the current operation step and the corresponding spatial threshold, a spatial risk function for the current step is generated. Generate the corresponding dynamic risk function based on the spatial risk function. Then the dynamic risk function It can be represented as ; In the formula, For steps The time activation weight is a time-related factor. The function that reflects the steps At any moment The activity level can be obtained through modeling, such as represented as ; here, For steps Estimated execution center time; To control the duration of the risk impact of this step, it is usually combined with This suggests that the risks of a step emerge and diminish smoothly before and after its execution, rather than switching instantaneously.

[0020] The spatial risk function includes point-like risk sources (such as live terminals) and regional risk sources (such as high-altitude work surfaces).

[0021] When the risk source is a point-like structure, a radial attenuation kernel is used, represented as follows: ; When it is a regional risk source, a spatial indicator kernel is used, represented as follows: ; In the formula, Coordinates of the risk source center; To control the radius of risk impact; This is an indicator function.

[0022] Dynamic risk function It is determined by the currently activated step.

[0023] The perception and scheduling module is used to schedule and configure monitoring areas and perception resources in real time based on the dynamic risk heat map, and construct the current control zone; based on the generated control zone, it obtains the status information of the operators in real time.

[0024] In this embodiment, based on a dynamic risk function, mobile and fixed sensing resources are optimized for scheduling to form a focused current control area. And efficiently obtain personnel status information The monitoring area includes high-risk areas and areas of human activity as shown on the risk heat map, and the sensing resources consist of deployed mobile monitoring devices and nearby fixed monitoring devices. High-risk areas are defined based on risk thresholds, as shown below. ; The specific value of the risk threshold can be set according to the actual situation.

[0025] The personnel activity area is defined based on the real-time location of personnel. Specifically, a coverage area is generated centered on the location of all workers, based on real-time positioning. The controlled area is jointly determined by both parties, and is represented as... ; In the formula, The kernel density estimation function for personnel location; , These are the weighting coefficients; Generate thresholds for the controlled area.

[0026] The resulting control zone is further optimized based on its coverage quality and scheduling costs, with the goal of maximizing coverage quality and minimizing scheduling costs. The coverage quality of the control zone is represented as... ; In the formula, For a set of sensing devices; For each sensing device; This is the current controlled area; The perceptual efficiency of point x; The configuration of each device, including its location, angle, and operating mode, determines its perception efficiency for a point x in space. This efficiency value ranges from 0 to 1 and is related to its resolution, distance, and angle. The volume (or area) of the control zone effectively covered by at least one device is calculated, and the product term represents the probability that no device can cover it.

[0027] Scheduling costs This includes the device's energy consumption, communication overhead, and mode switching overhead.

[0028] The final optimization objective is then expressed as: ; in, This represents the cost penalty coefficient. The optimization process is solved within each scheduling cycle, thereby optimizing scheduling and configuration strategies in a timely manner and reducing energy consumption and costs.

[0029] Suppose we have M mobile sensing devices and N fixed sensing devices (such as fixed cameras and sensors). Each device has a configurable sensing range (field of view, coverage area) and sensing capabilities (video, gas, temperature, etc.). The scheduling objective is to select a set of devices and their configurations (such as the location, orientation, and sensor mode of mobile devices) to maximize coverage of the controlled area, while considering resource constraints (such as the mobile energy consumption and communication bandwidth of mobile devices).

[0030] Under the dispatched sensing network, personnel status information is acquired in real time. Each For each worker, the data is represented by a vector containing information such as position, posture, tool usage status, and physiological indicators. The data is fused from multiple sources using Kalman filtering or deep neural networks to improve the accuracy of state estimation. In this embodiment, the worker's position can be obtained using coordinates obtained through positioning technologies such as UWB. Behavioral status can be identified through video analysis of mobile devices, recognizing actions such as walking, climbing, and tool operation. Physiological status data, such as heart rate and body temperature, is obtained through wearable devices, while environmental status data, such as gas concentration, temperature, and humidity, is obtained from sensors integrated into the mobile device.

[0031] The risk prediction module is used to analyze the coupling risk coefficient between status information and work stage based on work elements, obtain the predicted risk coefficient based on the coupling risk coefficient and work intention, and output intervention measures based on the risk prediction.

[0032] In this embodiment, the dynamic deviation between the "current work step requirements" and the "actual state of the personnel" is analyzed, and coupled with future work intentions, a risk coefficient is predicted. The coupled risk coefficient is calculated comprehensively based on the spatial deviation, behavioral deviation, tool deviation, and physiological deviation of the current personnel j, and can be expressed as follows: In the formula, This is the weight vector; This is the deviation vector.

[0033] In this embodiment, spatial deviation is the deviation of the personnel's position from the current step's permissible threshold; behavioral deviation is the semantic distance between the personnel's actions and the expected actions of the step, which can be identified and determined through visual recognition of the personnel's actions and the expected actions of the step; tool (equipment) deviation is the proportion of missing or incorrect tools, which can be calculated based on visual or RFID detection; physiological deviation is the personnel's heart rate variability, the normalized difference between the heart rate variability (HRV) and the baseline value.

[0034] Based on the intended purpose of the task, risk prediction is performed, and the predicted risk coefficient is expressed as follows: ; In the formula, This is the weight vector; For predicting personnel state transitions, sequence models such as LSTM can be used to make predictions based on recent state sequences; The coupling coefficient is the intended coupling coefficient. The step-switching risk gain represents the increment of the baseline risk when switching from the current step to the next step (which can be learned from historical data). This is the current step; For the next step; Activate when approaching a step switch; This describes the status of personnel as they approach the next step. This allows the system to issue warnings before personnel perform incorrect steps or prematurely enter the next risk area.

[0035] In this embodiment, the intervention measures are dynamically selected and output based on the risk coefficient, predicted risk, and specific deviation type.

[0036] Option 2 A method for intelligent dynamic safety management and control of power outsourcing operations is provided, which is applied to the aforementioned intelligent dynamic safety management and control system for power outsourcing operations, as shown in the attached figure. Figure 2 As shown, it includes the following steps: S1: Obtain current task information and decompose the task information into four-tuple task elements; generate dynamic risk heatmaps for each task step based on the task elements.

[0037] In this embodiment, by acquiring the electronic work order, operation ticket, and other work information of the current task, and using natural language processing and knowledge graph technology, the work information is decomposed into structured work elements, including a work step sequence S, a work space sequence L, a work resource set R, and a step risk label set H. Based on the work step sequence and the risk label corresponding to each step, combined with the spatial region of the step and the risk kernel function, a dynamic risk heatmap that changes over time is generated. This heatmap is composed of a dynamic risk function. It means that, among them, Let x represent time and x represent spatial coordinates.

[0038] S2, based on a dynamic risk heat map, combines the current work steps with the scheduling and configuration of monitoring areas and sensing resources to construct the current control zone; and obtains the status information of the workers in real time.

[0039] In this embodiment, the current risk salience area and personnel activity area are calculated based on the dynamic risk function at the current moment and the real-time personnel location information. The risk salience area and personnel activity area are then weighted and merged to generate the current control zone, which is the area that needs to be monitored closely at the moment.

[0040] Based on the controlled area and available sensing resources (including mobile and fixed devices), a sensing resource scheduling strategy is derived to maximize coverage quality and minimize scheduling costs within the controlled area. Following this strategy, sensing resources (such as the location, angle, and operating mode of mobile devices, and the operating mode of fixed devices) are configured.

[0041] By using a configured sensing resource network, multi-source data (video, location, gas concentration, etc.) from the work site are collected in real time, and the data is fused to obtain the status information vector of the workers. .

[0042] S3 calculates the coupling risk coefficient in real time based on the work elements, current steps, and personnel status information, and calculates the predicted risk coefficient in combination with the work intention to perform risk prediction.

[0043] In this embodiment, for each worker, based on the requirements of the current work step (obtained from the work elements) and the actual status information, a deviation vector is calculated in multiple dimensions such as space, behavior, tools, and physiology. Using the deviation vector and a predefined weight vector, the real-time coupling risk coefficient of each worker is calculated.

[0044] Based on the current work steps and the intention of the next work steps (step sequence), combined with the historical state sequence, a predictive model is used to predict the state deviation and risk coefficient in the future period, and the predicted risk coefficient is obtained.

[0045] S4, based on risk prediction, output operational intervention measures according to intervention rules.

[0046] In this embodiment, a set of intervention measures is generated according to predetermined intervention rules based on the real-time risk coefficient, the predicted risk coefficient, and the specific type of deviation. The intervention measures include the intervention target, the intervention method, and the intervention intensity.

[0047] The intervention rule refers to providing prompts or interventions based on the determined intervention intensity level, as expressed as follows: ; In the formula, For dynamic risk coefficients; To predict risks; This refers to the intensity threshold. Specifically, alerts or alarms can be issued to specific workers through means such as sound, light, vibration, and displays. If necessary, mandatory intervention can be carried out through the linkage control system (such as power outage or lockout).

[0048] This embodiment also includes a dynamic update step. By monitoring the work progress, when a work step changes, the current step index is updated, and the elements and risk heatmap of the new step are retrieved again from the work parsing step. The perception scheduling step and the risk prediction and intervention step are executed cyclically until the work is completed. Through the cyclical execution of the above steps, dynamic and intelligent safety management and control of the entire outsourced work process is achieved. Each step is dynamically adjusted based on the output of the previous step and real-time data, forming a closed-loop control system.

[0049] In this embodiment, a dynamic risk heatmap with spatiotemporal linkage is generated based on work ticket parsing, enabling risk visualization to evolve with the work steps. An optimized algorithm adaptively schedules sensing resources to form a dynamic control area focusing on key risks, achieving dynamic adjustment of the monitoring focus. A predictive mechanism is also introduced, analyzing the deviation between personnel's real-time status and work intentions to achieve a shift from post-event alarms to pre-event prevention. This solution effectively solves the persistent problems of large blind spots, shallow identification, and slow response in traditional monitoring, increasing monitoring coverage to over 95%, achieving centimeter-level risk identification and positioning accuracy, and shortening emergency response time to less than one minute. This significantly improves the intensity and scope of monitoring, effectively reduces regulatory costs, enhances overall management efficiency, and ensures the safety and reliability of outsourced operations.

[0050] The above descriptions are merely embodiments of the present invention, and common knowledge such as specific technical solutions and / or characteristics are not described in detail here. It should be noted that those skilled in the art can make various modifications and improvements without departing from the technical solutions of the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. An intelligent dynamic security management system for power outsourcing, characterized in that, Comprise: Job analysis module, for the obtained job information is decomposed into the current job elements, generate the dynamic risk heat map of the current job; The job elements include job steps, job space, job resources and job risks; Perception scheduling module, for real-time scheduling and configuration of monitoring area and perception resources according to the dynamic risk heat map, build the current control area; Based on the generated control area, the state information of the job personnel is obtained in real time; Risk prediction module, for analyzing the coupling risk coefficient of state information and job stage based on job elements, combining job intention to obtain prediction risk coefficient based on coupling risk coefficient, and outputting intervention measures according to risk prediction.

2. The intelligent dynamic security management system for power delegation according to claim 1, characterized in that: In the job analysis module, the key information is extracted from the electronic work ticket, and the information is mapped to four job elements to form the element set E, wherein, The job steps are represented as ; each step contains a description, a standard execution time and a set of predecessor steps; The working space is represented as ; each step is associated with a three-dimensional space region, then is the permission threshold for the step, ; is the coordinate of a point in the space; is a three-dimensional space region; , defines a polyhedral region; The job resource is represented as wherein, is a vector of tool list; is a vector of personal qualification matrix; is a vector of personal protective equipment list; The job risk is expressed as ; each .

3. The intelligent dynamic security management system for power delegation according to claim 2, characterized in that: generating a spatial risk function for the current step based on the current work step and the corresponding spatial threshold generating a corresponding dynamic risk function based on the spatial risk function is represented as ； wherein is the time activation weight of step ; is the time, x is the spatial coordinate.

4. The intelligent dynamic security management system for power delegation according to claim 3, characterized in that: The space risk function includes point risk source and area risk source; When it is a point risk source, radial attenuation kernel is adopted, which is expressed as ； When a regional risk source, a spatial indicator core is employed, denoted as ; wherein is the risk source center coordinate; is the control risk influence radius; is the indicator function.

5. The intelligent dynamic security management system for power delegation according to claim 1, characterized in that: In the perception scheduling module, the monitoring area is the displayed high risk area and personnel activity area, and the perception resource is the configured mobile monitoring device and the nearby fixed monitoring device; The high risk area is demarcated according to the risk threshold, and the personnel activity area is demarcated according to the real-time position of personnel.

6. The intelligent dynamic security management system for power delegation according to claim 5, characterized in that: The formed control area is also optimized and adjusted according to the control area coverage quality and scheduling cost; The control area coverage quality is expressed as ； wherein, is a set of sensing devices; is each sensing device; is a current control zone; is a sensing performance of point x; is a configuration for each device, including position, angle, and operating mode; The scheduling cost is the energy consumption, communication overhead and mode switching overhead of the equipment.

7. The intelligent dynamic security management system for power delegation operation according to claim 3, characterized in that: In the risk prediction module, the coupling risk coefficient is calculated by comprehensively considering the spatial deviation, behavior deviation, tool deviation and physiological deviation of the current personnel; The spatial deviation is the deviation of personnel position from the current step permission threshold; The behavior deviation is the semantic distance between personnel action and step expected action; The tool deviation is the proportion of missing or wrong tools; The physiological deviation is the heart rate variability of personnel.

8. The intelligent dynamic security management system for power delegation according to claim 7, characterized in that: The prediction risk coefficient is expressed as ； wherein, is a weight vector; is a bias vector; is an intent coupling coefficient; is a step switch risk gain; is a current step; is a next step; is activated in proximity to a step switch; is a state of the person in proximity to the next step.

9. A smart dynamic security management method for power outsourcing, characterized in that, The intelligent dynamic safety control system for power outsourcing operation according to any one of the above claims 1-8 comprises the following steps: S1, obtain the current job information, and decompose the job information into four job elements; According to the job elements, generate the dynamic risk heat map corresponding to each job step; S2, based on the dynamic risk heat map, combine the current job step scheduling and configuration of monitoring area and perception resources to build the current control area; And real-time acquisition of state information of job personnel; S3, according to the job elements, the current step and the personnel state information, the coupling risk coefficient is calculated in real time, and the prediction risk coefficient is calculated by combining the job intention, and the risk prediction is carried out; S4, according to the risk prediction, the intervention measures of the job are output according to the intervention rules.

10. The intelligent dynamic security management method for power outsourcing according to claim 9, characterized in that: The intervention rules are to prompt or intervene according to the judgment of intervention intensity level, which is expressed as ； wherein is a dynamic risk factor; is a predicted risk; is a strength threshold.

Images