A virtual reality-based safety training method, equipment, and medium for the power industry.

Abstract

This invention discloses a method, device, and medium for safety training in the power industry based on virtual reality, relating to the field of power simulation technology. The method includes: acquiring raw data of operational behavior, including timestamps, hand spatial trajectories, action triggering states, and power equipment numbers; generating action segment data based on the hand spatial trajectories and action triggering states, and constructing action chain data based on the timestamps and power equipment numbers; calculating and generating sequence deviation scores and path offset scores based on the action chain data and standard process data; generating evaluation score records based on the sequence deviation scores, path offset scores, and duration in the action segment data; and generating evaluation results containing step numbers, score values, and action offset markers based on the evaluation score records. This invention achieves automatic identification and deviation quantification evaluation of key action steps in power operation processes without the need for manual expert review or subjective scoring standards.

Description

Technical Field

[0001] This invention relates to the field of power simulation technology, specifically to a method, equipment, and medium for power industry safety training based on virtual reality. Background Technology

[0002] The application of virtual reality technology in safety training in the power industry can create a highly realistic working environment for trainees, enabling them to complete simulated drills of high-risk tasks such as switching operations, equipment inspections, and abnormal handling under non-real-world conditions.

[0003] Currently, the standardization training and evaluation of power operation procedures typically relies on human experts to analyze videos or data records step by step and judge whether the execution of each key step is standardized based on subjective experience. This process is not only time-consuming and labor-intensive, but also easily affected by the experience differences of the evaluators when dealing with continuous, cross-cutting, and multi-objective operations, resulting in a lack of uniformity and objectivity in the evaluation conclusions. Especially in large-scale training scenarios, it is difficult to form a data foundation that can be used for precise control and feedback loops, and there is a lack of systematic expression of the action structure and time logic in the operation process, making it difficult to support the quantitative modeling of deviations in training behavior.

[0004] Secondly, operational behavior data is often mixed with a large number of spatial and temporal deviations introduced by individual differences, environmental interference, or differences in equipment response, making it difficult to effectively separate the core action segments with operational intent. In scenarios with path overlap, positioning errors, or behavioral jumps, there is a lack of an expression mechanism that can correspond behavioral deviations to operational nodes one by one, and it is also impossible to construct a data expression method that includes scoring information and spatial markers. This makes it difficult to provide quantitative feedback on the compliance of specific action positions, thereby limiting the accurate positioning and visual recognition of behavioral deviation positions. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method, equipment, and medium for safety training in the power industry based on virtual reality.

[0006] A virtual reality-based safety training method for the power industry, the method comprising: S11: Obtain the raw data of the operation behavior, which includes timestamp, hand spatial trajectory, action triggering status and power equipment number; S12: Generate motion fragment data based on hand spatial trajectory and motion triggering state, and construct behavior chain data based on timestamp and power equipment number; S13: Calculate and generate sequence deviation score and path offset score based on behavioral chain data and standard process data; S14: Generate an evaluation score record based on the sequence deviation score, path offset score, and duration in the action segment data, and generate an evaluation result containing step number, score value, and action offset marker map based on the evaluation score record.

[0007] Furthermore, the steps to obtain the raw data of the operational behavior are as follows: S111: The spatial positioning device deployed in the simulation environment collects the three-dimensional position coordinates of the user's hand at each time stamp, and constructs the hand spatial trajectory data by combining the time stamp sequence; S112, by operating the button event detection module in the touch panel, obtain the user's action trigger status in each operation cycle, and record the corresponding power equipment number; S113 integrates the hand spatial trajectory, timestamp, action trigger status, and power equipment number to generate raw data of the operation behavior.

[0008] Furthermore, the steps for generating motion fragment data are as follows: S121, Based on the timestamps corresponding to each sampling point in the hand spatial trajectory, identify sampling point segments that are continuous in time but have significant spatial position changes, and organize them into continuous trajectory segment data; S122, Based on the start and end times recorded in the action trigger state, extract the trajectory segments that intersect with the time intervals in the continuous trajectory segment data, and generate an initial action segment set; S123: Remove trajectory segments that do not meet the minimum duration and sampling quantity requirements from the initial action segment set, and output the remaining trajectory segments as action segment data.

[0009] Furthermore, the steps for generating the initial set of action segments are as follows: S122.1 Extract all start and end time pairs recorded in the action trigger state and use them as a set of action time periods; S122.2, sequentially compare the start and end times of each continuous trajectory segment with the set of action time periods to determine whether there is any effective overlap; S122.3, Filter out all trajectory segments with overlapping time intervals and generate an initial action segment set.

[0010] Furthermore, the steps for constructing behavioral chain data based on timestamps and power equipment numbers are as follows: S124, Sort the action segment data according to the timestamp information to determine the time sequence of each action segment in the actual execution process; S125: Based on the power equipment number information, the time-series ordered action segments are mapped to the corresponding equipment identifiers, and the equipment number and time sequence position information are marked. S126 organizes all labeled action segments into an ordered structure in chronological order to generate behavior chain data.

[0011] Furthermore, the generation logic for the order deviation score and path offset score is as follows: S131, construct operation execution sequence data based on the power equipment number and triggering sequence corresponding to each action segment in the behavior chain data; S132, construct standard process sequence data based on the predefined equipment number sequence and operation sequence requirements in the standard process data; S133, Input the operation execution sequence data and the standard process sequence data into the edit distance algorithm to calculate the sequence deviation score; S134, based on the difference comparison of the three-dimensional position coordinates of all action segments in the behavior chain data and the standard process data, generates a path offset score.

[0012] Furthermore, the steps for generating path offset scores include: S134.1 Extract the three-dimensional position coordinates corresponding to all action segments in the behavior chain data and organize them into a trajectory comparison sequence according to the timestamp order; S134.2 Extract the three-dimensional spatial position points corresponding to each standard action in the standard process data, and arrange them in numerical order to generate a standard trajectory sequence; S134.3 calculates the difference between the trajectory comparison sequence and the standard trajectory sequence in both time and space dimensions, statistically obtains the average degree of offset, and outputs a path offset score.

[0013] Furthermore, the steps for generating evaluation results include: S141. Calculate the comprehensive score of each action segment based on the sequence deviation score, path offset score, and duration corresponding to each action segment. S142, bind the comprehensive score value with the corresponding action segment trigger sequence, organize them in chronological order, and generate an evaluation score record; S143, Based on the comprehensive score value of each action segment in the evaluation scoring record and the path offset score, determine the corresponding action offset marker status; S144, the step number, comprehensive score and action offset marker status of the action segment are structurally integrated to generate an evaluation result containing the step number, score and action offset marker map.

[0014] An electronic device, comprising: Memory, used to store computer software programs; A processor is used to read and execute the computer software program, thereby implementing any of the virtual reality-based power industry safety training methods described above.

[0015] A non-transitory computer-readable storage medium storing a computer software program, which, when executed by a processor, implements any one of the virtual reality-based safety training methods for the power industry.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention, through the standardized integration of timestamps, hand spatial trajectories, action trigger states, and power equipment numbers in the original data of operational behavior, can construct a training data structure with temporal consistency and spatial positioning correlation without the need for manual expert review. This supports the logical chain expression and execution sequence evaluation of each key action step in the power operation process, thereby improving the ability to quantify the degree of behavioral deviation in the operation process. In addition, based on the continuity of action trigger states and hand spatial trajectories, it can distinguish between valid operational behaviors and unexpected interference data. Combined with data normalization processing, it effectively eliminates the dimensional influence between different devices, providing a consistent measurement basis for behavior analysis and improving the applicability and comparability of the overall training data.

[0017] Furthermore, this invention constructs behavioral chain data containing timestamps and power equipment numbers, and combines sequence deviation scoring and path offset scoring to achieve multi-dimensional evaluation of the operation execution process. Based on this, it generates evaluation results containing step numbers, score values, and action offset marker maps, which can reflect the execution quality status based on the comprehensive score value of each action segment. By associating the action offset marker status with spatial position changes, it marks and visualizes potential abnormal locations in the operation process, thereby improving the spatial positioning capability of violations without the need for subjective scoring standards, and promoting the automation and standardization of the training and evaluation process in the power industry.

[0018] In summary, this invention enables the automatic identification and deviation quantification assessment of key action steps in the power operation process without the need for manual expert review and subjective scoring standards. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0020] Figure 1 This is a flowchart of a virtual reality-based safety training method for the power industry provided in Embodiment 1 of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, 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.

[0022] Example 1 Please see Figure 1 As shown in the figure, this embodiment discloses a safety training method for the power industry based on virtual reality, the method including: S11: Obtain the raw data of the operation behavior, which includes timestamp, hand spatial trajectory, action triggering status and power equipment number; In one specific embodiment, the system obtains raw data of operational behavior through the perception layer component in the virtual reality environment for power industry safety training; obtains the timestamp of each operation moment by calling the timeline interface of the virtual reality engine; and obtains the hand spatial trajectory, action triggering state, and power equipment number by combining the deployed spatial positioning device and touch feedback module.

[0023] Specifically, the steps to obtain the raw data of the operation behavior are as follows: S111: The spatial positioning device deployed in the simulation environment collects the three-dimensional position coordinates of the user's hand at each time stamp, and constructs the hand spatial trajectory data by combining the time stamp sequence; In one specific embodiment, the spatial positioning device captures the three-dimensional position coordinates of the user's hand in the virtual coordinate system at a preset sampling frequency; By linearly connecting the three-dimensional position coordinates at each time point, hand spatial trajectory data is constructed; Specifically, the calculation logic for constructing hand spatial trajectory data is as follows: First, calculate the differences between the three-dimensional position coordinates at the current timestamp and the three-dimensional position coordinates at the previous timestamp in the horizontal, vertical, and longitudinal directions. Calculate the squares of the differences in the three directions separately and sum them to obtain the sum of squares of the displacements; The square root of the sum of squared displacements is used to obtain the segmented displacement length of the current sampling period; The segmented displacement length is added to the total trajectory length accumulated in the previous cycle to generate hand spatial trajectory data. It should be noted that the sampling frequency is obtained through the hardware performance parameters of the spatial positioning device.

[0024] Preferably, the sampling frequency can be 90 Hz.

[0025] S112, by operating the button event detection module in the touch panel, obtain the user's action trigger status in each operation cycle, and record the corresponding power equipment number; In one specific embodiment, the button event detection module monitors the voltage level signal output by the touch panel in real time to obtain the user's action trigger state in each operation cycle; Record the power device number targeted by the current interaction. Specifically, the calculation logic for obtaining the action trigger state is as follows: Read the real-time voltage value of the touch panel during the current operation cycle; If the real-time voltage value is greater than or equal to the preset logic judgment threshold, the action trigger state will be marked as logic value 1. If the real-time voltage value is less than the preset logic judgment threshold, the action trigger state will be marked as logic value 0. It should be noted that the logic judgment threshold is obtained from the logic circuit specification of the touch panel.

[0026] Preferably, the logic threshold can be 3.3 volts.

[0027] S113, integrate the hand spatial trajectory, timestamp, action trigger status and power equipment number to generate raw data of operation behavior; In one specific embodiment, based on the timestamp, the hand spatial trajectory, action triggering state and power equipment number are mapped to a unified feature vector, and structural integration is performed to generate the original data of the operation behavior; Specifically, the calculation logic for generating the original data of the operation behavior is as follows: The hand spatial trajectory, action triggering state, and power equipment number are encapsulated into a preset data matrix in chronological order; Normalization is performed on each row vector in the data matrix. The normalization method is as follows: Divide each component value in the vector by the corresponding maximum feature parameter; The obtained component values ​​are the original data for the operation behavior; It should be noted that the maximum feature value parameter is obtained based on the preset coordinate range and total number of devices in the simulation environment.

[0028] The normalized component values ​​range from 0 to 1, with an example value of 0.85.

[0029] The present invention provides a virtual reality-based safety training method for the power industry, which has at least the following technical effects: By combining spatial positioning devices with timestamps to construct hand spatial trajectory data, high-granularity reconstruction of power operation actions is achieved; by using logic levels to determine the action trigger state, the user's interaction intentions among complex power equipment can be accurately captured, effectively avoiding accidental touches and misjudgments; the original operational behavior data generated by structural integration is normalized to eliminate the influence of dimensions, providing a standardized data source for subsequent accurate assessment of operational compliance and violation judgment; the higher the sampling frequency, the higher the fitting accuracy of the hand spatial trajectory data, thereby further improving the pertinence and effectiveness of safety training.

[0030] S12: Generate motion fragment data based on hand spatial trajectory and motion triggering state, and construct behavior chain data based on timestamp and power equipment number; In a specific embodiment, the system reads the previously collected hand spatial trajectory and recorded action triggering state from the memory buffer; performs segmentation operation using the displacement characteristics of the hand spatial trajectory and the time interval of the action triggering state to generate action segment data; and constructs behavior chain data by extracting the timestamp associated with the action segment data and combining it with the power equipment number corresponding to the operation target to perform logical concatenation.

[0031] Specifically, the steps for generating motion fragment data are as follows: S121, Based on the timestamps corresponding to each sampling point in the hand spatial trajectory, identify sampling point segments that are continuous in time but have significant spatial position changes, and organize them into continuous trajectory segment data; In one specific embodiment, the system parses the timestamp information carried by each sampling point in the hand's spatial trajectory; By calculating the displacement changes of sampling points corresponding to adjacent timestamps in three-dimensional space, sampling point segments that are continuous on the time axis and have significant spatial position changes are identified and organized into continuous trajectory segment data. Specifically, the calculation logic for identifying sampling points with significant changes in spatial location is as follows: First, extract the horizontal coordinates, vertical coordinates, and height coordinates of two adjacent sampling points; Calculate the squares of the differences in the horizontal coordinates, the vertical coordinates, and the height coordinates, and sum the three squared values ​​to obtain the sum of squared coordinates. The square root of the sum of squared coordinates is used to obtain the displacement values ​​of adjacent points; Divide the displacement value of adjacent points by the corresponding timestamp difference to obtain the instantaneous movement rate; If the instantaneous movement speed is greater than or equal to the preset motion significance threshold, the corresponding sampling point will be marked as a significant change point. Collect and organize the significant changes in timestamps into continuous trajectory segment data; It should be noted that the motion salience threshold is preset based on the average execution speed of the electrical operation action.

[0032] Preferably, the motion salience threshold can be selected as 0.05 meters per second.

[0033] S122, Based on the start and end times recorded in the action trigger state, extract the trajectory segments that intersect with the time intervals in the continuous trajectory segment data, and generate an initial action segment set; In one specific embodiment, the system reads the action trigger state data stream and identifies the physical feedback moment when the user presses or triggers the interactive switch; based on the start time and end time recorded in the action trigger state, it extracts the trajectory portion that overlaps with the continuous trajectory segment data on the time axis and generates an initial action segment set.

[0034] Specifically, the steps for generating the initial set of action segments are as follows: S122.1 Extract all start and end time pairs recorded in the action trigger state and use them as a set of action time periods; In one specific embodiment, the system retrieves the moment when the logical value changes from 0 to 1 in the action trigger state as the start time, and records the moment when the logical value returns to 0 as the end time. A set of action time periods is constructed by pairing each pair of adjacent start and end times.

[0035] S122.2, sequentially compare the start and end times of each continuous trajectory segment with the set of action time periods to determine whether there is any effective overlap; In a specific embodiment, each data item in the continuous trajectory segment data is traversed to obtain the start time and stop time of the trajectory segment; the start time and stop time of the trajectory segment are compared with each time interval in the action time period set to determine whether there is a valid overlap.

[0036] S122.3, Filter out all trajectory segments with overlapping time intervals and generate an initial action segment set.

[0037] In one specific embodiment, the system retains all trajectory information of the detected overlapping time windows, stores it in temporary storage space, and generates an initial action segment set.

[0038] S123, filter out trajectory segments that do not meet the minimum duration and sampling number requirements from the initial action segment set, and output the remaining trajectory segments as action segment data; In one specific embodiment, the system performs data quality filtering on the initial set of action segments, filtering out trajectory segments that do not meet the minimum duration and sampling quantity requirements, and outputting the remaining trajectory segments that meet the conditions as action segment data; Specifically, the calculation logic for filtering out trajectory segments that do not meet the requirements is as follows: Calculate the difference between the end time and the start time of each trajectory segment in the initial action segment set to obtain the duration value; Count the total number of timestamps contained in the trajectory segment to obtain the sample size value; If the duration value is greater than or equal to the preset lower time threshold and the sample size value is greater than or equal to the preset lower quantity threshold, then the trajectory segment is determined to be valid and retained. If the duration value is less than the lower time limit threshold, or the sample size value is less than the lower quantity limit threshold, then a removal operation is performed; It should be noted that the time limit threshold and the quantity limit threshold are used to exclude interference from unconscious shaking or misoperation by trainees in the virtual environment.

[0039] Preferably, the lower limit threshold for time can be 0.2 seconds, and the lower limit threshold for quantity can be 10.

[0040] Specifically, the steps for constructing behavioral chain data based on timestamps and power equipment numbers are as follows: S124, Sort the action segment data according to the timestamp information to determine the time sequence of each action segment in the actual execution process; In one specific embodiment, the system reads the first timestamp information of each trajectory in the action segment data; by sorting the first timestamps in ascending order, the temporal order of each action segment in the actual execution process is determined.

[0041] S125: Based on the power equipment number information, the time-series ordered action segments are mapped to the corresponding equipment identifiers, and the equipment number and time sequence position information are marked. In one specific embodiment, the system retrieves the power equipment number corresponding to each action segment, matches the corresponding physical equipment attributes in the equipment database of the virtual simulation scene, and maps them to the corresponding equipment identifier; in the data structure, an attribute field is added for each action segment, and the equipment number and the timing position information reflecting the operation sequence are marked.

[0042] S126: Organize all labeled action segments into an ordered structure in chronological order to generate behavioral chain data; In one specific embodiment, the system logically attaches the labeled action segments according to the numerical values ​​of the temporal position information from smallest to largest, generating behavior chain data; Specifically, the calculation logic for generating behavior chain data is as follows: the device number, geometric features of the hand spatial trajectory, and temporal position information of each action segment are encapsulated into a linked storage structure; the timestamp difference between two adjacent action segments is calculated to obtain the action interval duration; the action interval duration is stored as an association attribute between nodes in the linked storage structure to generate behavior chain data; it should be noted that the behavior chain data fully expresses the logical sequence of actions performed by the trainee on a specific power device.

[0043] The virtual reality-based safety training method for the power industry provided in this invention has at least the following technical effects: By recognizing the spatial change amplitude of the hand's spatial trajectory, it can effectively eliminate invalid dwell time during the operation process, and the generated continuous trajectory segment data can more accurately lock the effective operation area; by combining action trigger state execution with dual screening, it ensures that the generated action segment data has a clear operation intention and eliminates environmental noise interference; by constructing behavior chain data containing equipment attributes, it realizes digital modeling of the power operation process. When the sampling frequency is higher, the action connection recorded in the behavior chain data is smoother, thereby significantly improving the accuracy of identifying violations (such as skipping steps or misoperation).

[0044] S13: Calculate and generate sequence deviation score and path offset score based on behavioral chain data and standard process data; In one specific embodiment, the system extracts standard process data corresponding to the current training task from the database; compares the actual operation logic recorded in the behavior chain data with the preset specifications in the standard process data, and generates sequence deviation score and path offset score through quantitative calculation.

[0045] Specifically, the generation logic for order deviation scores and path offset scores is as follows: S131, construct operation execution sequence data based on the power equipment number and triggering sequence corresponding to each action segment in the behavior chain data; In one specific embodiment, the system traverses the behavior chain data and extracts the power equipment number associated with each action segment; according to the order of triggering timing, the power equipment numbers are arranged into a linear data structure to generate operation execution sequence data.

[0046] S132, construct standard process sequence data based on the predefined equipment number sequence and operation sequence requirements in the standard process data; In one specific embodiment, the system parses the compliant operation path for a specific power operation task in the standard process data; extracts the predefined equipment number sequence and operation sequence requirements, and organizes them into standard process sequence data.

[0047] S133, Input the operation execution sequence data and the standard process sequence data into the edit distance algorithm to calculate the sequence deviation score; In one specific embodiment, the system compares the degree of change of operation execution sequence data relative to standard process sequence data; Specifically, the calculation logic for generating the order bias score is as follows: The total number of insert, delete, and replace operations required to convert the operation execution sequence data into standard flow sequence data is calculated to obtain the edit distance value. The total number of power equipment numbers contained in the standard process sequence data is used as the standard sequence length value; Divide the edit distance value by the standard sequence length value to obtain the order deviation score; It should be noted that the order deviation score needs to be normalized. The normalization method is to limit the result of the above ratio to between 0 and 1. If the calculated result is greater than 1, the value is taken as 1. The sequence deviation score is obtained by calculating the degree of deviation of the operational logic. The larger the sequence deviation score, the less the trainer's operation steps conform to the standard specifications.

[0048] Preferably, the standard sequence length value can be obtained based on the specific number of switching operation steps.

[0049] S134, Based on the difference comparison of the three-dimensional position coordinates of all action segments in the behavior chain data and the standard process data, a path offset score is generated; In one specific embodiment, the system extracts the actual motion space information from the behavior chain data and compares it point by point with the theoretical spatial position in the standard process data; by statistically analyzing the deviation of spatial distance, a path offset score is generated.

[0050] Specifically, the steps for generating path offset scores include: S134.1 Extract the three-dimensional position coordinates corresponding to all action segments in the behavior chain data and organize them into a trajectory comparison sequence according to the timestamp order; In one specific embodiment, the system extracts the horizontal, vertical, and height coordinates of all trajectory points from the behavior chain data to form three-dimensional position coordinates; the above coordinates are arranged in order of timestamps to form a trajectory comparison sequence.

[0051] S134.2 Extract the three-dimensional spatial position points corresponding to each standard action in the standard process data, and arrange them in numerical order to generate a standard trajectory sequence; In one specific embodiment, the system reads the three-dimensional spatial location points of the power equipment interaction center points from standard process data; arranges them according to the operation order of the power equipment numbers to generate a standard trajectory sequence.

[0052] S134.3 calculates the difference between the trajectory comparison sequence and the standard trajectory sequence in both time and space dimensions, statistically obtains the average degree of offset, and outputs the path offset score; In one specific embodiment, the system performs dynamic time warping to spatiotemporally correlate points in the trajectory comparison sequence with points in the standard trajectory sequence; Specifically, the calculation logic for generating path offset scores is as follows: calculate the three-dimensional spatial distance between each sampling point in the trajectory comparison sequence and the corresponding standard point in the standard trajectory sequence to obtain the single-point offset distance; The calculation method for single-point offset distance is as follows: calculate the squares of the differences in the horizontal coordinates, the vertical coordinates, and the height coordinates of the two points, sum them up, and then take the square root of the summation result; sum the single-point offset distances corresponding to all sampling points to obtain the total offset distance value; divide the total offset distance value by the total number of sampling points in the trajectory comparison sequence to obtain the average offset degree, and output it as the path offset score; it should be noted that the calculation of the average offset degree takes into account the geometric deviation of the path. When the path offset score is larger, it means that the trainee's operation trajectory is more unstable or the positioning is less accurate.

[0053] Preferably, the unit for path offset scoring can be set to meters.

[0054] S14: Generate an evaluation score record based on the sequence deviation score, path offset score, and duration in the action segment data, and generate an evaluation result containing step number, score value, and action offset marker map based on the evaluation score record; In a specific embodiment, the system retrieves the sequence deviation score and path offset score from the aforementioned steps, and extracts the duration corresponding to each action segment from the action segment data; it uses a preset comprehensive evaluation operator to perform a fusion operation on the sequence deviation score, path offset score, and duration to generate an evaluation score record; by parsing the data features in the evaluation score record and combining them with the rendering engine of the virtual scene, it finally generates an evaluation result containing the step number, score value, and action offset marker map.

[0055] Specifically, the steps to generate evaluation results include: S141. Calculate the comprehensive score of each action segment based on the sequence deviation score, path offset score, and duration corresponding to each action segment. In one specific embodiment, the system extracts the corresponding sequence deviation score, path offset score, and duration for each action segment, performs multi-factor weighted evaluation, and calculates the comprehensive score value of the action segment. Specifically, the calculation logic for the overall score of this action segment is as follows: First, calculate the difference between the first and second order deviation scores and multiply it by a preset order weight coefficient to obtain the first evaluation component. Next, the difference between the path offset score and the score is calculated and multiplied by a preset path weight coefficient to obtain the second evaluation component. The duration is normalized and multiplied by a preset time weighting coefficient to obtain the third evaluation component; Finally, the first evaluation component, the second evaluation component, and the third evaluation component are summed to obtain the comprehensive score. It should be noted that the sequence weight coefficient, path weight coefficient, and time weight coefficient are preset by the power training expert system, and the sum of the three is equal to 1. The method for normalizing the duration is to divide the duration by the preset duration quota in the standard process data. The value ranges from 0 to 1, and an example value can be 0.85. The smaller the sequence deviation score, the more compliant the operation sequence is with the standard, and the higher the calculated comprehensive score value.

[0056] Preferably, the order weight coefficient can be 0.4.

[0057] S142, bind the comprehensive score value with the corresponding action segment trigger sequence, organize them in chronological order, and generate an evaluation score record; In one specific embodiment, the system obtains the action segment trigger sequence associated with each action segment in the behavior chain data; attaches the calculated comprehensive score value as an attribute field to the corresponding action segment trigger sequence node; sorts all comprehensive score values ​​in ascending order according to the order of action segment trigger sequences, and generates an evaluation score record.

[0058] S143, Based on the comprehensive score value of each action segment in the evaluation scoring record and the path offset score, determine the corresponding action offset marker status; In one specific embodiment, the system traverses the evaluation and scoring records and extracts the comprehensive score value and path offset score associated with each action segment; The corresponding action offset marker state is determined by using a logical threshold judgment method; Specifically, the logic for determining the corresponding action offset marker state is as follows: If the overall score is less than the preset qualified score threshold, or the path deviation score is greater than or equal to the preset deviation warning threshold, then the corresponding action deviation marker status is determined to be a warning. If the overall score is greater than or equal to the qualified score threshold and the path deviation score is less than the deviation warning threshold, then the corresponding action deviation marker status is determined to be normal. It should be noted that the pass / fail scoring threshold and the deviation warning threshold are preset by the power safety training management system.

[0059] Preferably, the pass / fail score threshold can be 0.8.

[0060] S144, The step number, comprehensive score and action offset marker status of the action segment are structurally integrated to generate an evaluation result containing the step number, score and action offset marker map; In one specific embodiment, the system obtains the step number corresponding to each action segment and encapsulates it with the corresponding comprehensive score and action offset marker status. Specifically, the logic for generating an evaluation result containing the step number, score, and action offset marker map is as follows: the comprehensive score is mapped to a score value; a status identifier is superimposed on the trajectory projection point in the virtual three-dimensional space according to the action offset marker status, and it is converted into a two-dimensional plane action offset marker map; the step number, score, and action offset marker map are filled into the evaluation report template to generate the evaluation result.

[0061] The present invention provides a virtual reality-based safety training method for the power industry, which has at least the following technical effects: By comprehensively considering sequence deviation scores, path offset scores, and duration, the operational level of trainees can be evaluated in a multi-dimensional and quantitative manner, and the generated comprehensive score value scientifically reflects the quality of operation; by using the action offset marker status to determine the action offset marker diagram, the spatial visualization of violations is realized. When the path offset score is larger, the warning color block in the action offset marker diagram is more prominent, which helps trainees to intuitively discover the deviation position of the operation trajectory; the step number and score value are structurally integrated in the evaluation results, realizing accurate traceability and assessment of power operation processes, thereby improving the digital level and closed-loop management efficiency of power industry safety training.

[0062] Example 2 This embodiment discloses an electronic device, including: Memory, used to store computer software programs; A processor is used to read and execute the computer software program, thereby implementing any of the virtual reality-based power industry safety training methods described above.

[0063] Since the electronic device described in this embodiment is the one used to implement the virtual reality-based power industry safety training method in this application embodiment, those skilled in the art can understand the specific implementation method and various variations of the electronic device in this embodiment based on the virtual reality-based power industry safety training method described in this application embodiment. Therefore, how the electronic device implements the method in this application embodiment will not be described in detail here. Any electronic device used by those skilled in the art to implement the virtual reality-based power industry safety training method in this application embodiment falls within the scope of protection of this application.

[0064] Example 3 This embodiment discloses a non-transitory computer-readable storage medium storing a computer software program, which, when executed by a processor, implements any one of the virtual reality-based power industry safety training methods described above.

[0065] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A safety training method for the power industry based on virtual reality, characterized in that, The method includes: S11: Obtain the raw data of the operation behavior, which includes timestamp, hand spatial trajectory, action triggering status and power equipment number; S12: Generate motion fragment data based on hand spatial trajectory and motion triggering state, and construct behavior chain data based on timestamp and power equipment number; S13: Calculate and generate sequence deviation score and path offset score based on behavioral chain data and standard process data; S14: Generate an evaluation score record based on the sequence deviation score, path offset score, and duration in the action segment data, and generate an evaluation result containing step number, score value, and action offset marker map based on the evaluation score record.

2. The power industry safety training method based on virtual reality according to claim 1, characterized in that, The steps to obtain the raw data of the operation behavior are as follows: S111: The spatial positioning device deployed in the simulation environment collects the three-dimensional position coordinates of the user's hand at each time stamp, and constructs the hand spatial trajectory data by combining the time stamp sequence; S112, by operating the button event detection module in the touch panel, obtain the user's action trigger status in each operation cycle, and record the corresponding power equipment number; S113 integrates the hand spatial trajectory, timestamp, action trigger status, and power equipment number to generate raw data of the operation behavior.

3. The power industry safety training method based on virtual reality according to claim 2, characterized in that, The steps to generate motion fragment data are as follows: S121, Based on the timestamps corresponding to each sampling point in the hand spatial trajectory, identify sampling point segments that are continuous in time but have significant spatial position changes, and organize them into continuous trajectory segment data; S122, Based on the start and end times recorded in the action trigger state, extract the trajectory segments that intersect with the time intervals in the continuous trajectory segment data, and generate an initial action segment set; S123: Remove trajectory segments that do not meet the minimum duration and sampling quantity requirements from the initial action segment set, and output the remaining trajectory segments as action segment data.

4. The power industry safety training method based on virtual reality according to claim 3, characterized in that, The steps to generate the initial set of action segments are as follows: S122.1 Extract all start and end time pairs recorded in the action trigger state and use them as a set of action time periods; S122.2, sequentially compare the start and end times of each continuous trajectory segment with the set of action time periods to determine whether there is any effective overlap; S122.3, Filter out all trajectory segments with overlapping time intervals and generate an initial action segment set.

5. A power industry safety training method based on virtual reality according to claim 4, characterized in that, The steps for constructing behavioral chain data based on timestamps and power equipment numbers are as follows: S124, Sort the action segment data according to the timestamp information to determine the time sequence of each action segment in the actual execution process; S125: Based on the power equipment number information, the time-series ordered action segments are mapped to the corresponding equipment identifiers, and the equipment number and time sequence position information are marked. S126 organizes all labeled action segments into an ordered structure in chronological order to generate behavior chain data.

6. A power industry safety training method based on virtual reality according to claim 5, characterized in that, The logic for generating order deviation scores and path offset scores is as follows: S131, construct operation execution sequence data based on the power equipment number and triggering sequence corresponding to each action segment in the behavior chain data; S132, construct standard process sequence data based on the predefined equipment number sequence and operation sequence requirements in the standard process data; S133, Input the operation execution sequence data and the standard process sequence data into the edit distance algorithm to calculate the sequence deviation score; S134, based on the difference comparison of the three-dimensional position coordinates of all action segments in the behavior chain data and the standard process data, generates a path offset score.

7. A power industry safety training method based on virtual reality according to claim 6, characterized in that, The steps for generating a path offset score include: S134.1 Extract the three-dimensional position coordinates corresponding to all action segments in the behavior chain data and organize them into a trajectory comparison sequence according to the timestamp order; S134.2 Extract the three-dimensional spatial position points corresponding to each standard action in the standard process data, and arrange them in numerical order to generate a standard trajectory sequence; S134.3 calculates the difference between the trajectory comparison sequence and the standard trajectory sequence in both time and space dimensions, statistically obtains the average degree of offset, and outputs a path offset score.

8. A power industry safety training method based on virtual reality according to claim 7, characterized in that, The steps to generate evaluation results include: S141. Calculate the comprehensive score of each action segment based on the sequence deviation score, path offset score, and duration corresponding to each action segment. S142, bind the comprehensive score value with the corresponding action segment trigger sequence, organize them in chronological order, and generate an evaluation score record; S143, Based on the comprehensive score value of each action segment in the evaluation scoring record and the path offset score, determine the corresponding action offset marker status; S144, the step number, comprehensive score and action offset marker status of the action segment are structurally integrated to generate an evaluation result containing the step number, score and action offset marker map.

9. An electronic device, characterized in that, include: Memory, used to store computer software programs; A processor is configured to read and execute the computer software program, thereby implementing the virtual reality-based power industry safety training method as described in any one of claims 1 to 8.

10. A non-transitory computer-readable storage medium, characterized in that, The storage medium stores a computer software program, which, when executed by a processor, implements the virtual reality-based safety training method for the power industry as described in any one of claims 1 to 8.

Images