A VR-based nuclear power plant human error prevention training system, platform and method

Abstract

The present invention belongs to the field of nuclear power human error prevention, and specifically relates to a VR-based nuclear power plant human error prevention training system, platform and method. It includes a hardware layer, a data layer and an application layer; the hardware layer outputs standardized data to the data layer, and the data layer outputs the generated index data to the application layer. The beneficial effects of the present invention are as follows: The functions and interfaces of each module of the present invention conform to the usage habits of nuclear power plant training trainees, ensuring that trainees can use the platform hall well through simple operation training, and can realize all-round trainee training such as scenario experience, human error prevention training, safety risk teaching, task operation evaluation, etc. The VR human error prevention scenario can be well integrated into each module, and trainees have a good sensory experience during the training process, ensuring the high efficiency of training.

Description

Technical Field

[0001] The present invention belongs to the field of nuclear power human error prevention, and particularly relates to a VR-based nuclear power plant human error prevention training system, platform and method. Background Art

[0002] To improve the operation safety of nuclear power plants, nuclear power takes "reducing unplanned shutdowns and zero fatalities" as the safety management goal. Currently, for the training related to human error prevention in high-risk operations, it mainly relies on lectures, PPTs, and watching animated videos. The knowledge transfer efficiency is very poor, with little effect, and even becomes a mere formality, leaving serious safety hazards for on-site operations. Due to factors such as the limited space of the nuclear power human factors laboratory and the complexity of some operations, the existing human factors training system cannot highly reproduce some nuclear power site task scenarios and corresponding human error traps.

[0003] The control of high-risk operations is mainly achieved through defensive human factors management, and its focus is on before the task execution. Therefore, it is necessary to further strengthen the human error prevention training for staff before the task execution. By conducting task scenario analysis, deduction or rehearsal, and carrying out human error prevention training based on the actual on-site scenario, if VR technology is applied to the specific task scenarios of nuclear power plants for human error prevention training, it can enable trainees to be on the scene, improve the teaching efficiency, and more proficiently master the application of human error prevention tools and related operation skills in high-risk operations. Summary of the Invention

[0004] The purpose of the present invention is to provide a VR-based nuclear power plant human error prevention training system and method. By combining a large number of human factors and safety event analyses, a targeted and practical human error trap library and safety risk library are developed, which can configure personalized courses for personnel with insufficient human error prevention skills and ineffective risk identification for training, so as to strengthen their human error prevention and safety skill levels, improve their psychological qualities, and thus reduce human errors.

[0005] The technical solution of the present invention is as follows: A VR-based nuclear power plant human error prevention training system includes a hardware layer, a data layer, and an application layer; the hardware layer outputs standardized data to the data layer, and the data layer outputs the generated index data to the application layer.

[0006] The hardware layer includes supporting hardware and VR devices, including a wearable virtual reality helmet, a VR eye tracker embedded in the helmet, a VR galvanic skin sensor, a VR training bench, a background server, a display device, and a data interface for interacting with an external resource platform. The corresponding VR devices collect behavior data, position data, face data, heart rate data, and eye movement data to form standardized data and output it to the data layer.

[0007] The data layer stores, aggregates, analyzes, and performs other heterogeneous fusion processing on the personnel behavior and physiological data collected by the hardware layer, establishes corresponding data evaluation models, converts them into data that can intuitively reflect key indicators such as personnel skill levels and emotional fluctuations, and outputs the indicator data to the application layer.

[0008] The application layer includes a front-end platform hall and a back-end management system. The front-end platform hall contains functional modules such as user login, VR scenario teaching, assessment, intelligent assistant, and operation playback; the back-end management system contains functional modules such as unit management, resource configuration, course management, personnel information management, and class management, and displays the relevant indicator data and management data during the process of carrying out human error prevention training on the front-end and back-end of the application layer.

[0009] A VR-based nuclear power plant human error prevention training platform includes a VR course resource library, a back-end management system, and a VR operation platform.

[0010] The front-end VR course resource library is based on major human factor / safety events that have occurred and high-risk operation scenarios, selects operation task scenarios that may lead to personnel casualties or human errors, analyzes the safety risk points and human error trap points therein, and forms a VR course resource library through on-site data collection and modeling. The VR scene display method in the VR course resource library is controlled by the back-end. The data generated during the VR virtual scene training is input into the back-end management system for data analysis and evaluation, and the data is further evaluated, verified, and optimized through the back-end feedback data.

[0011] The back-end management system controls the functions of the front-end, including VR display methods, personnel physiological information, and operation responses. By establishing a student ability evaluation model, it realizes the configuration of safety risk points and human error traps in the front-end VR scene courses. The data analyzed and evaluated by the back-end can be real-time fed back to the front-end for display.

[0012] The VR operation platform is the hardware carrier for the entire human error prevention training. Through the integration of various hardware, it realizes the real-time collection of data such as behavior data, location data, face data, heart rate data, and eye movement data, and feeds back the real-time calculation results to the operation end through the back-end analysis and calculation. All operation feedbacks in the operation platform will be real-time fed back to the back-end management system for calculation response, and the final results will be displayed on the front-end.

[0013] A VR-based nuclear power plant human error prevention training method includes the following steps:

[0014] Step 1: Build a VR scenario by setting up a spatial environment and a high-risk operation scenario at the nuclear power site, making it close to the real nuclear power site environment. Configure corresponding risk points and human error traps in the scenario to form various course resources, and import the resource libraries of various developed high-risk scenarios into the background management system. The library comprehensively covers practical operation task scenarios, possible risk points and related knowledge on-site, and human error traps that may induce human errors during the operation process;

[0015] Step 2: Import the course resource library into the background management system to form personalized course data;

[0016] Step 3: After the personalized course data configuration is completed, synchronize it to the VR operation platform. After selecting the corresponding course for experience training, interact with people and objects in the scenario in different roles, and quickly switch between different operation areas through the prompt of the plant mini-map or UI interface for operation;

[0017] Step 4: Through the VR eye tracker, analyze the eye movement states such as fixation, blinking, and saccade in the VR scenario, view data such as the fixation points, fixation rays, fixation trajectories of the subjects, and the eye movement heat maps of the interesting objects, and present them in a visual form;

[0018] Step 5: The galvanic skin response analysis module measures the sweat gland changes caused by mental activities and analyzes the related mental states as an indicator of cognitive effort;

[0019] Step 6: The data collected by the galvanic skin device and the eye tracker are synchronously transmitted to the background management system, analyzed and processed. According to the requirements, select specific human factor-related data segments, filter and extract the data results. After analysis by the corresponding evaluation model, the proficiency of practical skills is reflected in combination with the results of practical assessments.

[0020] The galvanic skin response analysis module in Step 5 quantitatively analyzes the degree of individual emotional arousal during the safety training process, including SC skin conductance data analysis, SCL (Tonic) phase-related galvanic skin signal analysis, and SCR event-related galvanic skin signal analysis.

[0021] The beneficial effects of the present invention are as follows: The functions and interfaces of each module of the present invention conform to the usage habits of nuclear power plant training trainees, ensuring that trainees can use the platform hall well through simple operation training, and realizing all-round trainee training such as scenario experience, human error prevention training, safety risk teaching, and task practical assessment. The VR human error prevention scenario can be well integrated into each module, and trainees have a good sensory experience during the training process, ensuring the high efficiency of training. It has the following major advantages:

[0022] 1) Human error prevention training function: By analyzing a large number of human error events, a relatively complete human error trap library and safety risk library have been formed. The VR scenarios are in line with on-site work practices, which can maximize the improvement of trainees' awareness of preventing human errors and safety literacy.

[0023] 2) Assessment function: The instructor sets corresponding personalized course content for trainees in the management platform. The human error trap points and safety risk points of the course can be randomly generated. After completing the study, trainees need to complete the assessment of corresponding theoretical knowledge points in the platform hall, play different job roles according to the task requirements or interact with other trainees in the same scenario to complete practical tasks. After the assessment, the system gives an all-round evaluation score and evaluation report covering awareness, behavior, psychological data, etc. The instructor can view the trainees' scores and abilities in the background in a timely manner, so as to understand the trainees' mastery of the course content of this period, etc.

[0024] 3) Competition function: Using cloud platform technology, through VR technology, complex environments are simulated and corresponding competition practical scenarios are set. The participants in the competition complete the operation of the same task within the same time. The system gives scores and ranks. This function can be used for skills competitions such as human error prevention in power plants to promote the improvement of personnel's corresponding skill levels.

[0025] 4) Comprehensive evaluation function: Combining the assessment results of trainees' theory, practice, and psychological indicators, the system will comprehensively calculate the overall level of trainees according to the set weight coefficients and evaluation models, and give the corresponding scores and completion status of each sub-item, forming an evaluation report.

[0026] The VR-based human error prevention training device has low requirements for the venue, can solve the problems existing in physical human factor laboratories, such as insufficient work task scenarios, single human error traps, and little challenge, etc., realize the diversity and flexibility of human factor safety teaching, complement physical human factor laboratories, improve the comprehensiveness of human error prevention training, and improve the efficiency of human error prevention training. Description of the Drawings

[0027] Figure 1 It is the architecture diagram of a VR-based nuclear power plant human error prevention training system provided by the present invention;

[0028] Figure 2 It is the training flow chart of the human factor comprehensive training platform in the implementation of the present invention. Detailed Embodiment

[0029] The present invention will be further described in detail below with reference to the drawings and specific embodiments.

[0030] The invention relates to a VR-based human error prevention training system and method for nuclear power plants, which utilizes virtual reality technology to carry out human factor safety visualization training and experiential training (experiencing scenarios such as electric shock, fire, high fall, mechanical injury, etc.), and conducts deduction and rehearsal of important tasks before starting work.

[0031] A VR-based human error prevention training system for nuclear power plants includes a hardware layer, a data layer, and an application layer.

[0032] Hardware layer: It covers supporting hardware and VR devices, including wearable virtual reality helmets, VR eye trackers embedded in the helmets, VR galvanic skin sensors, VR training gantries, back-end servers, display devices, and data interfaces for interacting with external resource platforms. The corresponding VR devices collect behavioral data, position data, face data, heart rate data, eye movement data, etc. and form standardized data for output to the data layer.

[0033] Data layer: Through heterogeneous fusion processing such as storage, aggregation, and parsing of the personnel behavior and physiological data collected by the hardware layer, a corresponding data evaluation model is established, which is converted into data that can directly reflect key indicators such as personnel skill levels and emotional fluctuations, and the indicator data is output to the application layer.

[0034] Application layer: It mainly includes a front-end platform hall and a back-end management system. The platform hall contains function modules such as user login, VR scene teaching, assessment, intelligent assistant, and operation playback for use by instructors; the back-end management system is mainly for instructors and contains function modules such as unit management, resource configuration, course management, personnel information management, and class management. Combining the data analysis of the data layer, the relevant indicator data and management data of the trainees during the human error prevention training process are displayed on the front-end and back-end of the application layer to achieve the purpose of real-time monitoring, feedback, and evaluation.

[0035] During the training implementation process, the VR platform hall is mainly used for trainee training, and the indicators of the assessment data generated after training will be synchronized to the back-end management system; the back-end management system is mainly used by instructors. Based on the content in the resource library, a human error trap and safety risk library configuration is formed based on human factor event analysis, corresponding courses are set for trainees personalized, and a comprehensive evaluation report of the trainees is formed, so as to ensure that the instructors have the maximum control over the ability and quality levels of the trainees.

[0036] As Figure 2 shown, a VR-based human error prevention training platform for nuclear power plants includes three parts: a VR course resource library, a back-end management system, and a VR operation platform.

[0037] Front-end VR Course Resource Library. It mainly selects operation task scenarios that may lead to casualties or human error based on major human factor / safety incidents and high-risk operation scenarios that have occurred, analyzes the safety risk points and human error trap points therein, and forms a VR course resource library through on-site data collection and modeling. The VR scene display mode in the VR course resource library is controlled by the background. The data generated by the trainees during the VR virtual scene training is input into the background management system for data analysis and evaluation, and then evaluated, verified, and optimized through the feedback data from the background.

[0038] Background Management System. It mainly controls the functions of the front-end, including VR display mode, personnel physiological information, operation response, etc. At the same time, by establishing a trainee ability evaluation model, it realizes the configuration of safety risk points and human error traps in the front-end VR scene courses. The data analyzed and evaluated by the background can be real-time fed back to the front-end display.

[0039] VR Operation Platform. The VR operation platform is mainly the hardware carrier for the entire human error prevention training. Through the integration of various hardware, it realizes the real-time collection of data such as behavior data, position data, face data, heart rate data, and eye movement data, and feeds back to the operation end in real-time through the analysis and calculation of the background. At the same time, all operation feedbacks in the operation platform will be real-time fed back to the background management system for calculation response, and finally the final results will be displayed at the front-end.

[0040] A VR-based human error prevention training method for nuclear power plants, including

[0041] Step 1: Build a realistic VR scenario by building a spatial environment and high-risk operation scenarios at the nuclear power plant site, close to the real environment of the nuclear power plant site. Configure the corresponding risk points and human error traps in the scenario to form various course resources, and import the resource libraries of various developed high-risk scenarios into the background management system. The library comprehensively covers practical operation task scenarios, possible risk points and related knowledge on-site, and human error traps that may induce personnel errors during the operation process.

[0042] Step 2: After the course resources are imported into the background management system, when compiling the course, the instructor can configure it specifically from the human error trap library and the safety risk library to form a personalized course. The constructed high-risk operation scenarios and the set human error traps can provide action goals, action feedback, and behavior evaluation for the trainees.

[0043] Step 3: After the personalized course data configuration is completed, it will be synchronized to the VR operation platform. After selecting the corresponding course for experience training, through VR technology, interact with people and objects in the scenario in different roles, and quickly switch to each operation area for operation through the prompt of the plant mini-map or UI interface.

[0044] Step 4: During the operation process, through the VR eye tracker, the eye movement states such as fixation, blink, and saccade in the VR scene can be analyzed, and data such as the subject's fixation point, fixation ray, fixation trajectory, and eye movement heat map of the interesting object can be viewed and presented in a visual form.

[0045] Step 5: The galvanic skin response analysis module measures the sweat gland changes caused by mental activities and analyzes the related mental states as an indicator of cognitive effort. The galvanic skin response analysis module can quantitatively analyze the degree of individual emotional arousal during the safety training process, including SC skin conductance data analysis, SCL (Tonic) phase-related galvanic skin signal analysis, and SCR event-related galvanic skin signal analysis.

[0046] Step 6: The data collected by the galvanic skin device and the eye tracker are synchronously transmitted to the background management system for analysis and processing. According to the requirements, specific human factor-related data segments are selected, and meaningful data results are filtered and extracted. After being analyzed by the corresponding evaluation model, combined with the results of the practical operation assessment, to a certain extent, it can reflect the proficiency of practical skills and verify the rationality of the course setting through multiple trainings.

[0047] The present invention proposes a complete VR human factor operation safety training platform for nuclear power plants, including a hardware bench and a background management system, and combines the practical operation scenario course resources of nuclear power plant operation and maintenance personnel to simulate a virtual interaction scenario of nuclear power site operation, so as to realize a new virtual-real combination anti-human error training mode and method of "platform + resource".

Claims

1. A VR-based training method for preventing human error in nuclear power plants, characterized in that, It includes the following steps: Step 1: Build a VR scenario by setting up a spatial environment and a high-risk operation scenario at a nuclear power site, which is close to the real nuclear power site environment. Configure corresponding risk points and human error traps in the scenario to form various course resources, and import the resource libraries of various developed high-risk scenarios into the background management system. The library comprehensively covers practical operation task scenarios, possible risk points and related knowledge on site, and human error traps that may induce human errors during the operation process; Step 2: Import the course resource library into the background management system to form personalized course data; Step 3: After the personalized course data is configured, synchronize it to the VR operation platform. After selecting the corresponding course for experience training, interact with people and objects in the scenario in different roles, and quickly switch between different operation areas through the prompt of the plant mini-map or UI interface for operation; Step 4: Through a VR eye tracker, analyze the fixation, blink, and saccade states in the VR scenario, and view the fixation points, fixation rays, fixation trajectories of the test subjects, and the eye movement heat map of the interesting objects, which are presented in a visual form; Step 5: The galvanic skin response analysis module measures the sweat gland changes caused by mental activities and analyzes the related mental states as an indicator of cognitive effort; The galvanic skin response analysis module in Step 5 quantitatively analyzes the degree of emotional arousal of an individual during the safety training process, including SC skin conductance data analysis, SCL phase-related galvanic skin signal analysis, and SCR event-related galvanic skin signal analysis; Step 6: The data collected by the galvanic skin device and the eye tracker are synchronously transmitted to the background management system for analysis and processing. Select human factor-related data segments according to requirements, filter and extract the data results, and after analysis by the corresponding evaluation model, reflect the proficiency of practical skills in combination with the results of practical assessments.

Images