Immersive teaching and training system based on safety accident prevention

Abstract

The invention discloses an immersive teaching and training system based on safety accident prevention, and relates to the technical field of safety teaching and training. The system comprises a teaching and training management module, a scene experience module, a behavior supervision module and an effect evaluation module. The teaching and training management module constructs a cross-scene accident chain through multi-source data acquisition and parameter adaptation and adopts a Bayesian network model to quantify risks; the scene experience module is connected with the function scene unit through a standardized interface, and realizes multi-scene dynamic linkage based on accident chain logic; the behavior supervision module synchronously collects physiological and operation data, establishes a correlation model and executes hierarchical intervention; and the effect evaluation module integrates the multi-dimensional data to generate an evaluation report, and reversely optimizes teaching and training parameters. According to the system, through the technical means of multi-module cooperation, a whole-process closed loop of data acquisition, risk assessment, resource allocation, behavior supervision and effect optimization is realized, the practical training performance, accuracy and systematicness are improved, and effective support is provided for safety accident prevention.

Description

Technical Field

[0001] This invention relates to the field of safety education and training technology, specifically to an immersive education and training system based on safety accident prevention. Background Technology

[0002] In scenarios such as machining workshops and construction sites, safety training is a crucial link in accident prevention. In existing technologies, safety training equipment often exists independently in single-function scenarios. For example, mechanical accident simulation platforms can only reproduce a certain type of mechanical injury (such as roller entanglement or rolling mill crushing), simulated fire extinguishing experiences are limited to the operation of equipment in fixed fire scenarios, and comprehensive injury experiences only demonstrate a specific protective measure (such as falling helmets or electric shock).

[0003] However, traditional training methods have significant drawbacks: training scenarios are isolated, failing to replicate the chain reactions of accidents in real-world scenarios such as "mechanical injury leading to fire, and fire causing comprehensive injuries"; supervision of employee operations relies heavily on post-event manual recording, lacking real-time correlation analysis between physiological states (such as heart rate and stress levels) and violations, resulting in delayed interventions; and training effectiveness evaluation is limited to scoring individual items, failing to create a closed-loop process of "data collection - risk assessment - resource allocation - behavior supervision - effect optimization," ultimately leading to a disconnect between training content and actual risk characteristics, and limited improvement in employees' safety awareness and standardized operating skills. In summary, there is an urgent need for an immersive training system that enables dynamic multi-scenario linkage, precise in-process behavior intervention, and data-driven optimization throughout the entire process to address the core issues of isolated scenarios, delayed intervention, and difficulty in quantifying effects in safety training.

[0004] In view of the above, this application is hereby submitted. Summary of the Invention

[0005] The purpose of this invention is to provide an immersive training system based on safety accident prevention to solve the problems mentioned in the background art.

[0006] To address the aforementioned technical problems, this invention provides an immersive education and training system based on safety accident prevention, comprising an education and training management module, a scenario experience module, a behavior monitoring module, and an effect evaluation module. The training and education management module includes an accident chain management submodule, a data integration submodule, and a resource scheduling unit. The accident chain management submodule includes an accident chain modeling unit and a risk quantification submodule. The accident chain modeling unit has a built-in Bayesian network model for managing accident propagation logic. The data integration submodule includes a multi-source data acquisition unit and a parameter adaptation submodule. The scenario experience module includes a central control unit and at least two functional scenario units. The central control unit includes an instruction execution subunit and a scenario scheduling subunit. The behavior supervision module includes a signal acquisition unit and an intervention management subunit. The signal acquisition unit includes a physiological characteristic acquisition subunit and an operational behavior monitoring subunit. The effect evaluation module includes a data aggregation unit and an evaluation report generation subunit. The data aggregation unit includes a fault identification data receiving subunit, a behavior standard data receiving subunit, a handling process data receiving subunit, and an operational proficiency data receiving subunit. The training management module synchronizes operational data from the scenario experience module, behavior supervision module, and effect evaluation module in real time via an information interaction protocol. It calculates risk levels using a Bayesian network model in the incident chain management submodule and sends scenario resource scheduling instructions to the scenario scheduling submodule of the scenario experience module via the resource scheduling unit, while simultaneously sending behavior correction instructions to the intervention management submodule of the behavior supervision module. The data integration submodule collects equipment operation and maintenance data, employee operation data, and environmental parameter data through a multi-source data acquisition unit. After establishing the correspondence between data and scenario parameters through the parameter adaptation submodule, the data is transmitted to the incident chain management submodule. The behavior supervision module acquires employee physiological and operational data through a signal acquisition unit, and the intervention management submodule performs supervision and intervention. The effect evaluation module integrates data from all modules through a data aggregation unit, generates a management evaluation report through an evaluation report generation submodule, and feeds it back to the training management module. This achieves full-process management of training from "data acquisition - risk assessment - resource allocation - behavior supervision - effect optimization." This constructs a full-process management system suitable for safe training, enabling data-driven risk assessment and resource scheduling through multi-module collaboration, strengthening behavior supervision and effect optimization, and improving the systematicness and accuracy of training management.

[0007] Furthermore, in the training management module, the multi-source data acquisition unit of the data integration submodule is used to collect historical fault records of the equipment management system, employee operation behavior logs, and environmental monitoring data of the training scenario; the parameter adaptation submodule converts the integrated data into simulated parameters of the functional scenario unit of the scenario experience module, establishes a dynamic adaptation relationship between "actual operation and maintenance data and simulated scenario parameters", and optimizes the adaptation rules based on historical data; through a data integration and parameter adaptation mechanism specifically for training management, the consistency between simulated scenarios and actual risks is ensured, providing data support for the precise management of training content and improving the relevance of training content.

[0008] Furthermore, in the scenario experience module, the instruction execution subunit of the central control unit receives scheduling instructions from the resource scheduling unit of the education and training management module, and the scenario scheduling subunit controls the start, switching, and stop of each functional scenario unit according to the instructions; the central control unit connects with each functional scenario unit through a standardized interface to realize the rapid combination and splitting of functional scenario units; through the standardized interface and scheduling mechanism, flexible management and reuse of education and training scenario resources are realized, reducing the management cost of equipment deployment and improving the utilization efficiency of education and training resources.

[0009] Furthermore, the physiological characteristic acquisition subunit of the signal acquisition unit in the behavior supervision module is used to acquire the employee's heart rate and skin conductance response data, and the operation behavior monitoring subunit is used to identify the employee's protective equipment wearing status and the compliance of operation steps; the intervention management subunit establishes a correlation analysis model between physiological characteristics and operation behavior, and when a preset combination of violation characteristics is detected, it sends a pause or prompt command to the central control unit of the scene experience module; through real-time monitoring and correlation analysis, it realizes proactive supervision and intervention of unsafe behaviors, upgrades post-event correction to in-event management, and improves the standardization of employee safety behaviors.

[0010] Furthermore, in the training management module, the accident chain modeling unit of the accident chain management submodule performs the following steps: S1, receiving integrated data output by the data integration submodule; S11, constructing an accident propagation chain containing elements of "equipment failure - personnel operation - environmental impact" based on historical accident cases and fault records in the integrated data; S12, calculating the probability of occurrence of each link in the accident propagation chain by calling the Bayesian network model through the risk quantification subunit; S2, transmitting the accident propagation chain and probability data to the resource scheduling unit; by constructing the accident propagation logic and quantifying risks through standardized steps, a scientific basis is provided for the management decisions of key training areas, and the practicality of training is improved.

[0011] Furthermore, the risk calculation formula for the Bayesian network model is as follows: , in, This indicates that a fault characteristic has been detected. At that time, the accident chain link The conditional probability of occurrence Indicates the link in the accident chain. Fault characteristics appear when it occurs The probability, Indicates the link in the accident chain. The basic probability of occurrence. It represents the total number of links in the accident chain; through quantitative formulas, it enables accurate calculation of accident risks, provides data support for the management and allocation of education and training resources, avoids the blindness of resource allocation, and improves the scientific nature of risk prevention and control.

[0012] Furthermore, the education and training management module also includes a cloud-edge collaborative management unit, which comprises a cloud data management subunit and an edge control subunit. The cloud data management subunit is used to store the integrated data of the data integration submodule and the incident chain data of the incident chain management submodule, and to perform calculations of the Bayesian network model. The edge control subunit is used to receive cloud instructions and control the real-time operation of the scene experience module. Through the cloud-edge collaborative architecture, centralized management of education and training data and local real-time control of scenes are realized, improving the system's response speed and data management efficiency, and strengthening the stability and scalability of the education and training system.

[0013] Furthermore, in the effectiveness evaluation module, the fault identification data receiving subunit of the data aggregation unit receives the fault judgment results from the data integration subunit; the behavior norms data receiving subunit receives the behavior evaluation data from the behavior supervision module; the handling process data receiving subunit receives the process completion data from the accident chain management subunit; and the operation proficiency data receiving subunit receives the operation score data from the scenario experience module. The evaluation report generation subunit generates an evaluation report containing employee skill gaps and scenario optimization suggestions based on the four types of data and sends it to the data integration subunit of the training management module. By integrating multi-dimensional data, a training effectiveness evaluation report is generated, providing a management basis for optimizing subsequent training plans, forming a continuous improvement training management mechanism, and improving the closed-loop management level of safety training.

[0014] Compared with the prior art, the beneficial effects of the present invention are: Leveraging cross-scenario accident chain modeling and dynamic linkage technology, the practicality of training is significantly enhanced. By constructing a cross-scenario accident chain of "mechanical injury - fire - comprehensive injury," single-function scenario units (mechanical accident simulation platform, simulated fire extinguishing experience, and comprehensive injury experience) are linked according to the logic of real accident propagation. Scenario linkages are triggered based on real-time employee operation data (e.g., exceeding the time limit for mechanical operation triggers a fire scenario, or incorrect fire extinguishing operation triggers an electric shock scenario), breaking the limitations of isolated scenarios in traditional training and reproducing the chain reaction process of accidents. This technology allows employees to intuitively perceive the chain consequences of improper operation, significantly improving their awareness of complex risks and emergency response capabilities, and solving the problem of the disconnect between traditional training and actual accident scenarios.

[0015] By leveraging a dual-dimensional correlation analysis of physiological and behavioral factors and a tiered intervention technology, precise and proactive operational supervision is achieved. Through simultaneous collection of employee physiological data (heart rate, skin conductance) and operational behavior data (wearing of protective equipment, operational procedures), a "physiological state-operational behavior" correlation model is established. This model accurately distinguishes different types of violations, such as "nervous misoperation" and "habitual violations," and implements tiered interventions (Level 1: voice prompts; Level 2: scenario-based pause; Level 3: relearning of standardized procedures). This technology breaks through the traditional passive supervision model that relies solely on operational results, shifting the intervention timing from "post-event correction" to "in-event guidance," significantly improving the efficiency of correcting violations and reinforcing employees' habitual standardized operation.

[0016] By leveraging data-driven, end-to-end closed-loop optimization technology, we can continuously improve the quality of training and education. By integrating four types of data—fault identification, behavioral norms, handling procedures, and operational proficiency—we generate assessment reports that include suggestions for addressing skill gaps and scenario optimization. These reports then drive dynamic adjustments to scenario parameters (such as increasing the probability of triggering protective scenarios) and resource allocation strategies (such as prioritizing training for high-risk scenarios). This technology forms a complete closed loop of "data collection - risk assessment - experiential training - effectiveness optimization," ensuring that training content continuously adapts to actual risk characteristics and employee skill gaps. This addresses the problem of traditional training content being fixed and lacking dynamic improvement, enabling iterative upgrades to training effectiveness. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the principle of an immersive education and training system based on safety accident prevention. Figure 2 A flowchart of an immersive training system based on safety incident prevention; Figure 3 This is a schematic diagram of the mechanical accident simulation platform in this embodiment; Figure 4 This is a schematic diagram of the operation interface of the mechanical accident simulation platform in this embodiment; Figure 5 This is a schematic diagram of the simulated fire extinguishing experience unit in this embodiment; Figure 6 This is a schematic diagram of the operation interface of the simulated fire extinguishing experience unit in this embodiment; Figure 7 This is a schematic diagram of the integrated harm experience unit in this embodiment; Figure 8 This is a schematic diagram of the operation interface of the integrated injury experience unit in this embodiment. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figures 1 to 8 This invention provides a technical solution: an immersive training system based on safety accident prevention, applicable to safety training in machining workshops and construction sites. Through the collaboration of training management modules, scenario experience modules, behavior supervision modules, and effect evaluation modules, it integrates three core scenarios: a mechanical accident simulation platform, simulated fire extinguishing experience, and comprehensive injury experience. This achieves closed-loop management of the entire process from "data collection - risk assessment - resource allocation - behavior supervision - effect optimization," solving problems such as isolated scenarios, delayed intervention, and difficulty in quantifying effects in traditional training.

[0020] I. System Overall Architecture: This system comprises four main modules, each interacting with data via industrial Ethernet: Training Management Module: Includes an accident chain management submodule (with a built-in Bayesian network model), a data integration submodule (multi-source data acquisition and parameter adaptation), a resource scheduling unit, and a cloud-edge collaborative management unit, responsible for data integration, risk modeling, and resource scheduling; Scenario Experience Module: Includes a central control unit and three functional scenario units (mechanical accident simulation platform, simulated fire extinguishing experience unit, and comprehensive injury experience unit), enabling rapid combination and parameter adjustment through standardized interfaces; Behavior Supervision Module: Includes a signal acquisition unit (physiological characteristic acquisition subunit and operational behavior monitoring subunit) and an intervention management subunit, monitoring and intervening in employee operations in real time; Effectiveness Evaluation Module: Includes a data aggregation unit (four types of data receiving subunits) and an evaluation report generation subunit, generating multi-dimensional evaluation reports and optimizing the training process in reverse.

[0021] The following explanation uses a machine processing workshop (with a small construction site) as an example. This scenario presents high-frequency risks such as "injuries from machine rollers," "fires during hot work," and "injuries from falling objects due to unfastened safety helmets." The issue of employees not following proper protective procedures is prominent. The training area is 30 square meters. 2 Safety awareness needs to be improved through lightweight and precise education and training.

[0022] II. Education and Training Management Module: The education and training management module is the central hub of the system, responsible for data integration, incident chain modeling, risk quantification and resource scheduling. The operation process is divided into four steps: data collection and parameter adaptation, incident chain modeling and risk quantification, resource scheduling and cloud-edge collaborative management.

[0023] Step 1: Data Acquisition and Parameter Adaptation: The multi-source data acquisition unit collects historical injury cases, equipment operating parameters, and employee operation data from the mechanical accident simulation platform, simulated fire extinguishing experience, and comprehensive injury experience. The parameter adaptation subunit establishes a dynamic mapping between "actual data" and "scenario parameters," providing data support for risk modeling. Specific technical methods are as follows: 1. Multi-source data acquisition: Mechanical accident simulation platform: Collects historical data on roller entanglement injuries (averaging 3 per month, triggered by hand approaching the roller for 0.5 seconds), mill crush injuries (50N pressure threshold triggering simulated injury), sharp puncture injuries (30N triggering force), sharp blade cuts (20N blade pressure), and high-temperature burns (60℃ temperature triggering feedback); Simulated fire extinguishing experience: Collects records of handling 5 fires per year during hot work operations (average handling time 12 seconds) and 2 electrical fires per year (40% of fire extinguishers were the wrong type); Comprehensive injury experience: Collects data on violations such as 40% failure to fasten safety helmets, 35% failure to wear safety shoes, electric shock experience with 8mA current triggering numbness, and 30N clamping force from a press; Employee data: Records employee length of service, number of violations (such as operating machinery without protective gloves), and training history.

[0024] 2. Parameter Adaptation Rules: Mechanical Accident Simulation Platform: "Roller entanglement injury 3 times per month" is mapped to "simulated rotation speed 800 r / min, trigger probability 30%"; "rolling mill crush injury pressure 50N" is mapped to "simulated pressure 50N, feedback intensity 80%"; Simulated Fire Extinguishing Experience: "Hot work fire handling time 12 seconds" is mapped to "flame simulator height 40cm, burning speed 0.6m / min"; "fire extinguisher incorrect type selection rate 40%" is mapped to "incorrect type trigger probability 40%"; Comprehensive Injury Experience: "safety helmet not fastened violation rate 40%" is mapped to "falling object impact force 30N"; "safety shoe improper wearing rate 35%" is mapped to "static pressure test pressure 5500N"; Random forest algorithm is used to optimize the mapping threshold, using accident data from the past 3 years as the training set to iteratively improve the matching degree between scenario parameters and actual risks.

[0025] Example: In this scenario, the mechanical accident simulation platform experiences 24 sharp punctures per year on average, and the parameter adaptation subunit maps this to "20% probability of triggering simulated sharp puncture, and the activation weight of this unit is 20% each time the experience is performed"; the simulated fire extinguishing experience involves 5 hot work fires per year on average, which is mapped to "flame simulator burning speed 0.6m / min, fire extinguishing time limit 15 seconds"; the comprehensive injury experience shows that 35% of safety shoes are worn improperly, which is mapped to "static pressure test pressure 5500N, triggering 'improper wearing' audio and visual warning when improperly worn".

[0026] Existing technologies often use fixed preset values ​​for scenario parameters. This step collects actual injury data from the three experience units and dynamically adapts it, making the simulated scenario highly consistent with real risks and significantly improving the relevance of education and training.

[0027] Step 2: Accident Chain Modeling and Risk Quantification: Based on integrated data, a cross-scenario accident chain of "mechanical injury - fire - comprehensive injury" is constructed. A Bayesian network model is used to calculate the probability of occurrence at each stage, quantifying the risk level and providing a basis for resource scheduling. Specific technical methods are as follows: 1. Accident chain construction: Chain 1 (Mechanical-Fire chain): Node (Machine accident simulation platform roller entanglement injury, hands not protected) → Node (Employees panicked and touched the simulated fire extinguishing equipment, causing misoperation) → Node (Electric shock risk from comprehensive injury experience); Chain 2 (Fire - Comprehensive Injury Chain): Node (Incorrect fire extinguisher type selected during simulated fire extinguishing experience) → Node (The fire spread to the machinery area, causing high-temperature burns) → Node (Employee was injured by falling object because he was not wearing a safety helmet).

[0028] 2. Bayesian network computation: based on nodes (Roller entanglement) as an example, prior probability (Monthly occurrence probability), conditional probability The formula for calculating the posterior probability is: ; Substitute the data to get It was determined to be high-risk.

[0029] Example: Mill crush damage (node) on a mechanical accident simulation platform Prior probability conditional probability Substituting into the formula, we get the posterior probability. If a risk level is identified as high, the corresponding training resources will be allocated preferentially.

[0030] Existing accident chains are mostly based on single-scenario analysis. This step constructs cross-scenario accident chains and quantifies risks using Bayesian formulas, making risk assessment more scientific and resource scheduling more precise.

[0031] Step 3: Resource Scheduling: Based on the risk quantification results, send resource scheduling instructions to the three functional units of the scenario experience module to prioritize the training time and sequence for high-risk scenarios. Specific technical methods are as follows: 1. Time Allocation: High-risk scenarios (roller entanglement, hot work fires, falling objects from safety helmets): Training time ( (Number of employee-related violations); Medium-risk scenarios (sharp punctures, electrical fires, electric shock experiences): Minutes; Low-risk scenarios (cut by a sharp blade, dormitory fire, hand caught in a press): minute.

[0032] 2. Scene sequence: The training sequence for new employees is "Comprehensive injury experience unit (basic protection) → Mechanical accident simulation platform (mechanical risk) → Simulated fire extinguishing experience unit (fire risk)", forming a logical closed loop of "protection-operation-emergency".

[0033] Example: A new employee has two records of mechanical operation violations. Resource scheduling unit instructions: Mechanical accident simulation platform (35 minutes, focusing on roller entanglement and mill crushing injuries) → Simulated fire extinguishing experience unit (30 minutes, focusing on hot work fires) → Comprehensive injury experience unit (20 minutes, retesting protection effectiveness).

[0034] Existing technologies for resource scheduling lack contextual relevance and employee characteristic adaptation. This step combines cross-scenario risks and employee violation records to improve both resource utilization efficiency and training effectiveness.

[0035] Step 4: Cloud-Edge Collaborative Management: The cloud-based data management subunit stores historical data and incident chain models from the three experience units, while the edge control subunit controls local scenarios in real time, achieving "data to the cloud, control decentralized." Specific technical methods are as follows: 1. Cloud Operation: Stores 5 years of injury data from the mechanical accident simulation platform, fire handling records from simulated fire extinguishing, and comprehensive injury protection experience data. Performs Bayesian model iteration every morning to update the prior probabilities of each node.

[0036] 2. Edge control: Receive scene parameter commands from the cloud (such as the rolling mill crush pressure of 50N in the mechanical accident simulation platform), drive the actuators of each experience unit locally (such as pressure sensors and flame simulators), with a response time of ≤50ms; when the network is interrupted, cache employee operation data (such as the number of mechanical injury triggers), and synchronize to the cloud after the network is restored.

[0037] Example: When the workshop network is interrupted due to equipment interference, the edge control subunit caches 10 operation records of mechanical accident simulation (including roller entanglement trigger time and pressure value), and completes synchronization within 1 minute after the network is restored.

[0038] Existing cloud-edge collaboration technologies are not well-suited for complex scenarios with multiple experience units. This step significantly improves system stability and response speed through scenario-specific data storage and edge control.

[0039] III. Scenario Experience Module: The scenario experience module reproduces real risk scenarios through three major functional units. The operation process is divided into two steps: scenario unit deployment and parameter configuration, and cross-scenario linkage.

[0040] Step 5: Scenario Unit Deployment and Parameter Configuration: The central control unit connects to the mechanical accident simulation platform, simulated fire extinguishing experience unit, and comprehensive injury experience unit through standardized interfaces to quickly configure the parameters of each unit and reproduce simulated scenarios that match actual risks. Specific technical methods are as follows: 1. Mechanical accident simulation platform: Connects to the central control unit via USB-C interface, and displays operation instructions on the touch screen; the roller entanglement unit simulates a speed of 800r / min, the mill crushing unit simulates a pressure of 50N, the sharp puncture unit simulates a trigger force of 30N, the sharp blade cut unit simulates a blade pressure of 20N, and the high temperature burn unit simulates a temperature of 60℃.

[0041] 2. Simulated Fire Extinguishing Experience Unit: The main body of the equipment integrates four types of fire extinguisher interfaces, and buttons are used to select the fire type (hot work, electrical fire, etc.); the height of the flame simulator is adjusted according to the type (40cm for hot work, 30cm for electrical fire), and a built-in smoke generator simulates the spread of fire.

[0042] 3. Comprehensive Injury Experience Unit: The comprehensive experience platform integrates a safety helmet falling object device (3m height), a safety shoe protection experience area (static pressure 5000N), a simulated electric shock experience area (current 8mA), and a pressure machine hand clamping experience area (clamping force 30N); the touch screen displays the operation steps, guiding employees to complete the entire process of "protection-operation-emergency".

[0043] Example: During training, the central control unit instructs the mechanical accident simulation platform to activate the roller entanglement and high-temperature burn units, the fire extinguishing simulation unit to activate the hot work fire scenario, and the comprehensive injury simulation unit to activate the helmet falling object and electric shock simulation area, forming a coordinated training group. Employees experience roller entanglement (800 r / min) on the mechanical accident simulation platform, triggering simulated injury feedback when not wearing protective gloves; they then enter the fire extinguishing simulation unit, select a foam fire extinguisher to extinguish an electrical fire (incorrect type), triggering a "fire spread" warning; finally, they experience helmet falling object (30N impact force) and electric shock (8mA current) in the comprehensive injury simulation unit, completing the entire training process.

[0044] Existing technology's experience units are mostly independent devices. This step enables rapid combination of multiple units through standardized interfaces, improving deployment efficiency and enhancing scenario interoperability.

[0045] Step 6: Cross-Scene Collaboration: Based on the accident chain model, the scene scheduling subunit triggers the linkage between the mechanical accident simulation platform, the simulated fire extinguishing experience unit, and the comprehensive injury experience unit to reproduce the accident chain reaction. Specific technical methods are as follows: 1. Linkage Trigger Logic: The mechanical accident simulation platform detects that the roller is caught in a tangled mess and the person has not evacuated in time (operation timeout of 5 seconds), triggering the "accidental contact with equipment causing fire" scenario in the simulated fire extinguishing experience unit (flame height increases by 10cm); the simulated fire extinguishing experience unit selects the wrong type of fire extinguisher (such as a foam fire extinguisher for electrical fires), triggering the "electric shock risk" scenario in the comprehensive injury experience unit (current increases from 8mA to 10mA), and at the same time activating the safety helmet falling device (impact force increases from 30N to 40N).

[0046] 2. Parameter transmission: The degree of injury from mechanical accidents (such as the trigger time of roller entanglement) determines the fire spread rate in simulated fire extinguishing, and the fire spread rate determines the impact force of falling objects and the intensity of electric shock current in the overall injury.

[0047] Example: An employee exceeds the timeout period by 8 seconds in the roller entanglement unit of the mechanical accident simulation platform. The scenario scheduling subunit triggers a fire scenario in the simulated fire extinguishing experience unit, increasing the flame height from 40cm to 50cm. The employee selects the wrong fire extinguisher (using dry powder to extinguish an electrical fire), triggering the electric shock experience area in the comprehensive injury experience unit. The current increases from 8mA to 10mA, and simultaneously, the safety helmet's falling device activates, increasing the impact force from 30N to 40N. The employee must complete the chain of operations—"mechanical evacuation - correct fire extinguishing - emergency protection"—in sequence to complete the training.

[0048] Existing technologies often involve fixed processes for scenario linkage. This step, however, is based on the dynamic linkage between the accident chain and real-time operations, significantly improving the practicality of the training and employees' risk awareness.

[0049] IV. Behavior Monitoring Module: The behavior monitoring module monitors employees' operational behavior and physiological state in the three experience units in real time. The operation process is divided into two steps: behavior data collection and correlation analysis and graded intervention.

[0050] Step 7: Behavioral Data Acquisition: The physiological characteristic acquisition subunit (wristband device) of the signal acquisition unit collects employee heart rate and skin conductance response; the operational behavior monitoring subunit (industrial camera, sensor) identifies the operating procedures and protective equipment wearing of each experience unit. Specific technical methods are as follows: 1. Mechanical accident simulation platform: Industrial cameras identify whether hands are inserted into the roller entanglement area (with protective gloves on) and whether the emergency stop button is pressed; the physiological characteristic acquisition subunit monitors the employee's heart rate changes during operation (resting heart rate 70 beats / minute, rising to 95 beats / minute during operation).

[0051] 2. Simulated fire extinguishing experience unit: Sensors identify whether the fire extinguisher type selection is correct and whether the spray angle is ≥45°; the physiological characteristic acquisition subunit monitors the heart rate during fire extinguishing (up to 110 beats / minute when tense).

[0052] 3. Comprehensive injury experience unit: Industrial camera identifies whether the safety helmet chin strap is fastened and whether the safety shoes are worn correctly; sensor identifies the time of separation during electric shock experience (standard ≤3 seconds) and whether the operating gestures during the pressure machine hand clamping experience are correct.

[0053] Example: Employee B was operating a mechanical accident simulation platform without wearing protective gloves (recognized by the camera). His hand was inserted into the roller entanglement area for 0.8 seconds, triggering a simulated injury. His heart rate increased from 75 beats / minute to 100 beats / minute, and his skin conductance response was 0.7μS. In the comprehensive injury experience unit, the safety helmet chin strap was not fastened, triggering a "not wearing properly" prompt during the falling object experience. The safety shoes were worn properly, and no violation prompt was triggered when the static pressure test pressure was 5500N.

[0054] Existing behavioral monitoring technologies are mostly limited to single scenarios. This step covers physiological and operational monitoring across three major experience units, resulting in more comprehensive data and more accurate violation identification.

[0055] Step 8: Correlation Analysis and Tiered Intervention: The intervention management subunit establishes a "physiological-operational" correlation model to identify violation types (stress-induced misoperation, habitual violation) and implements tiered intervention. Specific technical methods are as follows: 1. Correlation Model Construction: Based on Heart Rate Increase Skin conductance Number of violations Operation time Based on the following characteristics, the classification rule is: Stress-related misoperations: times / minute , Seconds; Habitual violations: times / minute , Second.

[0056] 2. Tiered intervention implementation: Level 1 intervention (nervous misoperation): Voice prompt "Please check protective equipment before operating"; Level 2 intervention (habitual violation): Pause the scenario and play the standard operation video; Level 3 intervention (serious violation): Force relearn the relevant standards and pass the theoretical assessment before experiencing it again.

[0057] Example: Employee B's operation on the mechanical accident simulation platform was an example of misoperation due to stress. times / minute This triggers Level 1 intervention, with a voice prompt saying "Please wear protective gloves before operating"; not fastening the helmet properly in the comprehensive injury experience unit constitutes a habitual violation. times / minute This triggers a level-two intervention, pausing the scene and playing a video demonstrating proper helmet wearing procedures.

[0058] Existing intervention technologies lack grading and causal analysis. This step uses a correlation model to achieve precise grading intervention, resulting in more significant corrective effects.

[0059] V. Effectiveness Evaluation Module: The effectiveness evaluation module integrates data from the entire process, generates multi-dimensional evaluation reports, and optimizes the training process in reverse. The process consists of two steps: data collection and report generation, and parameter optimization feedback.

[0060] Step 9: Data Aggregation and Report Generation: The four sub-units of the data aggregation unit receive fault identification data from the training and education management module, behavioral norms data from the behavior supervision module, handling process data from the accident chain management sub-module, and operational proficiency data from the scenario experience module, respectively, and integrate them into a unified dataset; the evaluation report generation sub-unit generates a report containing capability shortcomings and optimization suggestions. Specific technical methods are as follows: 1. Data Summary: Fault Identification Data: Accuracy rate of injury type identification on the mechanical accident simulation platform (e.g., 80% identification rate of roller entanglement injury); Behavioral Norms Data: Compliance rate of protective equipment (e.g., 60% safety helmet fastening rate); Handling Process Data: Completeness of fire extinguishing steps (e.g., 70% accuracy rate of fire extinguishing steps during hot work); Operational Proficiency Data: Mechanical injury trigger time (e.g., 0.8 seconds trigger time for roller entanglement injury).

[0061] 2. Report generation: Calculate the overall score by weight (fault identification 25%, behavioral norms 30%, handling process 25%, and operational proficiency 20%), mark shortcomings (such as behavioral norms score of 60 points), and propose scenario optimization suggestions (such as adding protective exercises to improve the overall harm experience).

[0062] Example: Employee B's overall score: Fault identification 80 points, behavioral norms 60 points, handling procedures 70 points, operational proficiency 75 points, total score 71 points. Report recommendations: Add specific practice on safety helmets and safety shoes to the comprehensive injury simulation unit, and mandate checks on the wearing of protective equipment before each training session; add mandatory triggering logic for wearing protective gloves to the mechanical accident simulation platform.

[0063] The evaluation of existing technologies lacks multi-dimensional integration and specific recommendations. The report generated in this step is highly instructive and has a significant closed-loop optimization effect.

[0064] Step 10: Parameter Optimization Feedback: The evaluation report generation sub-unit sends optimization suggestions to the training and education management module, driving parameter optimization and resource scheduling adjustments across the three experience units. Specific technical methods are as follows: 1. Feedback Optimization: To address shortcomings in behavioral norms, the duration of the protective experience in the comprehensive injury experience unit has been increased (from 20 minutes to 30 minutes), and the trigger probability of the safety helmet falling object device has been increased (from 50% to 70%). To address insufficient fault identification, the trigger probability of injury types in the mechanical accident simulation platform has been increased (from 20% to 30%), and employees are required to identify the injury type before operating the system.

[0065] 2. Resource scheduling update: Prioritize training in the comprehensive injury simulation unit to ensure that employees' protective skills meet the standards before entering the mechanical accident simulation platform and the simulated fire extinguishing experience unit.

[0066] Example: Based on the shortcomings in the behavioral norms of multiple employees, the training management module instructs the integrated injury experience unit to increase the probability of the helmet falling object device triggering from 50% to 70%, and the static pressure test pressure of safety shoes from 5000N to 5500N; the resource scheduling unit raises the training priority of this unit to the highest level, and new employees must complete the training of this unit and meet the standards before they can enter other scenarios.

[0067] Existing technologies for improving education and training lack data-driven approaches. This step involves reverse optimization through evaluation reports, leading to continuous improvement in the quality of education and training.

[0068] In summary, this system overcomes the bottlenecks of traditional safety training, such as "isolated scenarios, delayed intervention, and difficulty in quantifying effects," through deep collaboration between a mechanical accident simulation platform, a simulated fire extinguishing experience unit, and a comprehensive injury experience unit. This is achieved by combining cross-scenario accident chain modeling and cloud-edge collaboration within the training management module, physiological-operational correlation intervention within the behavior monitoring module, and closed-loop optimization within the effect evaluation module. Compared to existing technologies, it achieves dynamic multi-scenario linkage (real-time operation triggering based on accident chains), precise hierarchical intervention (physiological-operational correlation model), and data-driven full-process optimization (parameter adjustment via evaluation reports). This significantly enhances the practicality and systematic nature of safety training, providing an effective solution for accident prevention in mixed scenarios of mechanical processing and construction.

Claims

1. An immersive training system based on safety accident prevention, characterized in that: It includes modules for education and training management, scenario experience, behavior monitoring, and effectiveness evaluation; The training and education management module includes an accident chain management submodule, a data integration submodule, and a resource scheduling unit. The accident chain management submodule includes an accident chain modeling unit and a risk quantification submodule. The accident chain modeling unit has a built-in Bayesian network model for managing accident propagation logic. The data integration submodule includes a multi-source data acquisition unit and a parameter adaptation submodule. The scene experience module includes a central control unit and at least two functional scene units. The central control unit includes an instruction execution subunit and a scene scheduling subunit. The behavior monitoring module includes a signal acquisition unit and an intervention management subunit. The signal acquisition unit includes a physiological characteristic acquisition subunit and an operational behavior monitoring subunit. The effectiveness evaluation module includes a data aggregation unit and an evaluation report generation subunit. The data aggregation unit includes a fault identification data receiving subunit, a behavior standard data receiving subunit, a handling process data receiving subunit, and an operation proficiency data receiving subunit. The training and education management module synchronizes the operational data of the scenario experience module, behavior supervision module, and effect evaluation module in real time through the information interaction protocol. It calculates the risk level through the Bayesian network model of the accident chain management submodule and sends scenario resource scheduling instructions to the scenario scheduling subunit of the scenario experience module through the resource scheduling unit. At the same time, it sends behavior correction instructions to the intervention management subunit of the behavior supervision module. The data integration submodule collects equipment operation and maintenance data, employee operation data, and environmental parameter data through the multi-source data acquisition unit. After the parameter adaptation subunit establishes the correspondence between the data and scenario parameters, it transmits the data to the accident chain management submodule. The behavior monitoring module acquires employees' physiological and operational data through the signal acquisition unit, and the intervention management sub-unit performs monitoring and intervention. The effect evaluation module integrates the data from each module through the data aggregation unit, and the evaluation report generation sub-unit generates a management evaluation report and feeds it back to the training management module, realizing the full-process management of training from "data acquisition - risk assessment - resource allocation - behavior monitoring - effect optimization".

2. The immersive training system based on safety accident prevention as described in claim 1, characterized in that: In the training and education management module, the multi-source data acquisition unit of the data integration submodule is used to collect historical fault records of the equipment management system, employee operation behavior logs, and environmental monitoring data of the training and education scenario; the parameter adaptation submodule converts the integrated data into the simulation parameters of the functional scenario unit of the scenario experience module, establishes a dynamic adaptation relationship of "actual operation and maintenance data - simulated scenario parameters", and optimizes the adaptation rules based on historical data.

3. The immersive training system based on safety accident prevention as described in claim 1, characterized in that: In the scenario experience module, the instruction execution subunit of the central control unit receives the scheduling instructions from the resource scheduling unit of the education and training management module. The scenario scheduling subunit controls the start, switching and stopping of each functional scenario unit according to the instructions. The central control unit connects with each functional scenario unit through a standardized interface to realize the rapid combination and splitting of functional scenario units.

4. The immersive training system based on safety accident prevention as described in claim 1, characterized in that: The physiological characteristic acquisition subunit of the signal acquisition unit in the behavior supervision module is used to acquire the employee's heart rate and skin conductance response data, and the operation behavior monitoring subunit is used to identify the employee's protective equipment wearing status and the compliance of operation steps; the intervention management subunit establishes a correlation analysis model between physiological characteristics and operation behavior, and when a preset combination of violation characteristics is detected, it sends a pause or prompt command to the central control unit of the scene experience module.

5. The immersive training system based on safety accident prevention as described in claim 1, characterized in that: In the training and education management module, the accident chain modeling unit of the accident chain management submodule performs the following steps: S1, receiving integrated data output by the data integration submodule; S11, constructing an accident propagation chain containing elements of "equipment failure - personnel operation - environmental impact" based on historical accident cases and fault records in the integrated data; S12, calculating the probability of occurrence of each link in the accident propagation chain by calling the Bayesian network model through the risk quantification subunit; S2, transmitting the accident propagation chain and probability data to the resource scheduling unit.

6. The immersive training system based on safety accident prevention as described in claim 5, characterized in that: The risk calculation formula for the Bayesian network model is as follows: ； in, This indicates that a fault characteristic has been detected. At that time, the accident chain link The conditional probability of occurrence Indicates the link in the accident chain. Fault characteristics appear when it occurs The probability, Indicates the link in the accident chain. The basic probability of occurrence. This indicates the total number of links in the accident chain.

7. The immersive training system based on safety accident prevention as described in claim 1, characterized in that: The training and education management module also includes a cloud-edge collaborative management unit, which comprises a cloud data management subunit and an edge control subunit. The cloud data management subunit is used to store the integrated data of the data integration submodule and the incident chain data of the incident chain management submodule, and to perform calculations of the Bayesian network model. The edge control subunit is used to receive cloud instructions and control the real-time operation of the scene experience module.

8. The immersive training system based on safety accident prevention as described in claim 1, characterized in that: In the effectiveness evaluation module, the fault identification data receiving subunit of the data aggregation unit receives the fault judgment results from the data integration subunit; the behavior norms data receiving subunit receives the behavior evaluation data from the behavior supervision module; the handling process data receiving subunit receives the process completion data from the accident chain management subunit; and the operation proficiency data receiving subunit receives the operation score data from the scenario experience module. The evaluation report generation subunit generates an evaluation report containing employee skill gaps and scenario optimization suggestions based on the four types of data and sends it to the data integration subunit of the training management module.

Images