Industrial injury prevention training system

Abstract

The invention relates to the technical field of data processing, in particular to an industrial injury prevention training system which comprises a construction module used for obtaining environment data flow on a real work site and constructing a digital twinborn model corresponding to the real work site according to the environment data flow; the simulation module is used for performing industrial injury prevention training on the user through a digital twinborn model and collecting behavior data of the user in the industrial injury prevention training process; the analysis module is used for performing industrial injury analysis according to the behavior data and generating industrial injury early warning information; and the output module is used for generating safety credit points according to the industrial injury early warning information and historical training records, so that the industrial injury prevention training system can perform accurate risk assessment and real-time early warning according to specific behaviors in personal actual operation of the user.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically to a workplace injury prevention training system. Background Technology

[0002] Work injury prevention training is an important measure to protect workers' safety and health and reduce the occurrence of production accidents. Its effective implementation plays a key role in improving the inherent safety level of enterprises.

[0003] Existing workplace injury prevention training systems mostly employ offline centralized lectures or online learning platforms, conveying general safety regulations and operational knowledge to employees through methods such as playing safety education videos and answering safety knowledge quizzes. However, because the training content and assessment methods heavily rely on pre-set fixed teaching materials and exams, they cannot conduct dynamic and personalized risk assessments and feedback based on individual employees' real-time operational behaviors. This results in insufficient training relevance and makes it difficult to accurately identify and warn of specific safety risks for particular employees in actual work environments. Summary of the Invention

[0004] To address the technical problem that existing work injury prevention training systems cannot achieve targeted risk assessment and early warning, this application provides a work injury prevention training system.

[0005] The work injury prevention training system provided in this application adopts the following technical solution: A workplace injury prevention training system includes: The module is used to acquire environmental data streams from the real work site and build a digital twin model corresponding to the real work site based on the environmental data streams. The simulation module is used to conduct workplace injury prevention training for users through a digital twin model, and to collect user behavior data during the workplace injury prevention training process; The analysis module is used to perform work injury analysis based on behavioral data and generate work injury early warning information; The output module is used to generate safety credit scores based on work injury early warning information and historical training records.

[0006] Furthermore, the steps of acquiring environmental data streams from the actual workplace and constructing a digital twin model corresponding to the actual workplace based on the environmental data streams include: By deploying each sensor network point on a real work site, environmental data streams are acquired at each sensor network point. Based on the edge computing nodes embedded in each sensor network point, each environmental data stream is processed to generate multiple environmental risk entropies. Based on multiple environmental risk entropies, an environmental risk topology map of the actual workplace is generated. According to the environmental risk entropy in the environmental risk topology, a grid resolution division threshold is set to redivide the preset three-dimensional network of the real work site, and a multi-resolution three-dimensional grid is obtained; The environmental risk entropy of each grid unit in the multi-resolution three-dimensional grid is determined, a preset texture-physical behavior model is loaded on the grid unit with the environmental risk entropy higher than the preset environmental risk entropy, and a preset texture-simplified physical model is loaded on the grid unit with the environmental risk entropy lower than the preset environmental risk entropy, and a digital twin model is obtained.

[0007] Further, based on the edge computing nodes embedded in each sensor network point, the steps of generating a plurality of environmental risk entropies include: The corresponding environmental data stream is normalized by each edge computing node to generate a plurality of environmental parameter vectors; The fluctuation amplitudes of each environmental parameter vector in each environmental dimension within a sliding time window are calculated to obtain a plurality of environmental fluctuation characteristics; After calculating the uncertainty measures of each edge computing node in each environmental dimension according to the environmental fluctuation characteristics, the uncertainty measures in each environmental dimension are fused to generate the environmental risk entropy corresponding to each edge computing node.

[0008] Further, the user is trained for work injury prevention through the digital twin model, and the steps of collecting the behavior data of the user during the work injury prevention training include: The user is presented with a virtual scene through the digital twin model; Based on the real-time operation of the user in the virtual scene, the digital twin model is simulated to obtain the environmental state changes and device interaction feedback caused by the real-time operation; The decision actions of the user under the environmental state changes and device interaction feedback are collected; The decision actions are quantified to obtain the behavior data.

[0009] Further, the steps of generating work injury warning information according to the behavior data include: The behavior data is feature extracted to obtain a behavior feature vector; The behavior feature vector is calculated by a preset safety evaluation model to obtain a safety deviation; The safety deviation is input into a preset risk evaluation rule for risk matching to obtain a risk type and a work injury risk level; The work injury warning information is generated according to the risk type and the work injury risk level.

[0010] Further, the steps of generating a safety credit score according to the work injury warning information and the historical training record include: The work injury early warning information and the historical training record are subjected to data quantification processing to obtain safety evaluation data; Based on the safety evaluation data, the work injury early warning information and the historical training record are subjected to weight distribution to obtain a weighted evaluation value; According to the weighted evaluation value, a score of a user in a user group to which the user belongs is calculated to obtain a relative safety score; The relative safety score is mapped to a preset integral range to obtain a safety credit score.

[0011] Further, the work injury early warning information includes a warning number and a warning level, and the historical training record includes a training score and a training completion degree. The step of subjecting the work injury early warning information and the historical training record to data quantification processing to obtain safety evaluation data includes: The warning number, the warning level, the training score and the training completion degree are subjected to unified dimension processing through a preset quantification index system to obtain a quantification index set; Each quantification index in the quantification index set is subjected to weighted fusion to generate a comprehensive safety evaluation value; The comprehensive safety evaluation value is mapped to a preset safety evaluation range to obtain safety evaluation data.

[0012] The beneficial effects achieved are: The application provides a work injury prevention training system, which comprises a construction module, a simulation module, an analysis module and an output module. The construction module is used to acquire environmental data flow on a real work site and construct a digital twin model corresponding to the real work site according to the environmental data flow. The simulation module is used to perform work injury prevention training on a user through the digital twin model and collect behavior data of the user in the work injury prevention training process. The analysis module is used to perform work injury analysis according to the behavior data and generate work injury early warning information. The output module is used to generate a safety credit score according to the work injury early warning information and historical training record.

[0013] That is, in the present application, the environment data stream of the real work site is obtained and the digital twin model is constructed by the construction module, thereby creating a highly simulated virtual training environment, which makes the training scene closely correspond to the real working conditions, then the user is trained for injury prevention in the digital twin model by the simulation module, and the behavior data of the user is collected in real time, which can reflect the specific operation and reaction of the user in the virtual training environment, providing a basis for personalized analysis, then the analysis module analyzes the injury according to the behavior data, generates injury warning information, the analysis process directly targets the real-time operation behavior of the user, realizes the transformation from general evaluation to individual risk assessment, and finally the output module generates safety credit points according to the injury warning information and historical training records, the safety credit points comprehensively consider the current injury warning and historical performance, forming a dynamic and targeted feedback mechanism, so that the injury prevention training system can accurately assess the risk and give real-time warning according to the specific behavior of the user in the actual operation. BRIEF DESCRIPTION OF DRAWINGS

[0014] Figure 1 is a schematic diagram of a module of an injury prevention training system of the present application; Figure 2 is a schematic diagram of the step flow of constructing a digital twin model of the present application; Figure 3 is a schematic diagram of the step flow of simulating training and collecting user behavior data by a digital twin model of the present application; Figure 4 is a schematic diagram of the step flow of generating injury warning information according to behavior data of the present application; Figure 5 is a schematic diagram of the step flow of generating safety credit points of the present application.

[0015] Explanation of the reference signs: 10, construction module; 20, simulation module; 30, analysis module; 40, output module. DETAILED DESCRIPTION

[0016] The following will be described in conjunction with the accompanying Figures 1-5 The present application will be further described in detail.

[0017] In the description of the present application, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and therefore cannot be understood as indicating or implying that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application. In addition, the terms "first", "second", "third" are only for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0018] In the description of the present application, it should be noted that unless specifically defined and limited, the terms "mounting", "connecting", "connecting" should be understood broadly, for example, it can be fixedly connected, or it can be detachably connected, or integrally connected; it can be mechanically connected, or it can be electrically connected; it can be directly connected, or it can be indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0019] The embodiment of the present application discloses a work injury prevention training system.

[0020] Please refer to Figure 1 The embodiment of the present application discloses a work injury prevention training system, which comprises: The construction module 10 is used for acquiring an environmental data stream on a real work site, and constructing a digital twin model corresponding to the real work site according to the environmental data stream; the simulation module 20 is used for conducting work injury prevention training on a user through the digital twin model, and collecting behavior data of the user in the work injury prevention training process; the analysis module 30 is used for conducting work injury analysis according to the behavior data, and generating work injury early warning information; and the output module 40 is used for generating safety credit points according to the work injury early warning information and historical training records.

[0021] The work injury prevention training system disclosed by the embodiment establishes a dynamic evaluation system capable of truly reflecting individual operation differences, converts the real work site into a digital twin model through the construction module 10, provides a highly realistic interactive basis for work injury prevention training, and keeps the virtual environment used for work injury prevention evaluation consistent with the real work site of the actual working condition. The simulation module 20 conducts work injury prevention training on the user on the digital twin model and synchronously records the behavior data of the user participating in the training, captures the individualized action habits and judgment logic exhibited by the user in the simulation operation, and provides a data basis for subsequent training analysis of the individual. The analysis module 30 generates work injury early warning information through analysis of the behavior data, realizes the conversion of abstract safety specifications into quantitative evaluation of specific operation behaviors of each user, and thus focuses the risk assessment on the individual actual operation level. The output module 40 generates safety credit points by comprehensively combining the early warning information and the historical records, forms a traceable and quantifiable individual safety performance index, and finally constructs a personalized work injury prevention mechanism based on data driving and capable of dynamically adjusting with the employee behavior, thereby effectively improving the accuracy of training evaluation and the pertinence of early warning.

[0022] In a feasible implementation manner, referring to Figure 2 The specific execution steps of the construction module include steps S11-S15. Step S11, through each sensor network point deployed on the real work site, the environmental data stream on each sensor network point is obtained.

[0023] By systematically deploying sensor network points composed of multiple types of environmental sensors in key areas of the real work site, each sensor network point can continuously collect environmental parameters such as temperature, humidity, illumination, and gas concentration at its location, forming a continuous environmental data stream. This deployment method ensures the spatial integrity and temporal continuity of environmental data stream collection, providing a comprehensive and real-time data basis for subsequent precise risk assessment. The most important effect is to establish a continuous and reliable data association between the digital twin model and the real work site, enabling the construction of a virtual environment and risk modeling based on real and dynamic on-site environmental data.

[0024] Step S12, based on the edge computing nodes embedded in each sensor network point, the respective environmental data streams are processed to generate multiple environmental risk entropies.

[0025] Based on the edge computing nodes embedded in each sensor network point, the environmental data stream collected from the corresponding sensor network point is processed to generate an environmental risk entropy corresponding to each sensor network point, thereby realizing real-time and localized calculation of environmental risk at the data collection source, thereby avoiding the direct transmission of a large amount of environmental data stream to the center system to reduce network bandwidth pressure and improve response speed.

[0026] In this step, the edge computing node performs real-time analysis on the environmental data stream such as temperature, humidity, illumination, and gas concentration, and fuses and converts multiple-dimensional environmental parameters into a single environmental risk entropy. This environmental risk entropy can quantitatively represent the local environmental instability and risk intensity of the area where each sensor network point is located, providing accurate and spatially distributed risk input data for generating the overall environmental risk topology map in the subsequent step, thereby ensuring that the basis for building the digital twin model is highly consistent with the dynamic risk conditions of the real work site.

[0027] Step S13, according to the multiple environmental risk entropies, an environmental risk topology map of the real work site is generated.

[0028] The environmental risk entropy calculated at each sensor network point is taken as a known data point by a spatial interpolation algorithm, and based on the corresponding spatial position coordinates, the environmental risk entropy estimation value of the area on the real work site where no sensor network point is deployed is calculated, so as to obtain the environmental risk status of each point in the continuous spatial range of the entire real work site, thereby generating a two-dimensional or three-dimensional environmental risk topology map covering the real work site completely, which is represented by different colors or numerical gradients to indicate the risk level, realizing the fusion and transformation of the originally isolated local risk information (i.e. environmental risk entropy) at each sensor network point into a whole, continuous and visual spatial risk distribution representation, thereby directly and accurately revealing the overall risk profile and spatial aggregation status of the high-risk area of the entire real work site, and providing a direct and accurate scientific basis for subsequent digital twin model for differential multi-resolution grid division.

[0029] Step S14, according to the environmental risk entropy in the environmental risk topology map, a grid resolution division threshold is set to redivide the preset three-dimensional network of the real work site, and a multi-resolution three-dimensional grid is obtained.

[0030] The distribution characteristics of the environmental risk entropy in the environmental risk topology map are determined through statistical analysis, for example, the environmental risk entropy is divided into different intervals according to the percentile and the corresponding resolution level is set, so as to establish a mapping rule between the environmental risk entropy and the grid fineness.

[0031] When redividing the preset three-dimensional grid, first, each grid cell in the preset three-dimensional grid is traversed, and the subdivision degree is dynamically adjusted according to the interval into which the corresponding environmental risk entropy falls. For the high-risk area grid with environmental risk entropy higher than the preset environmental risk entropy, recursive subdivision is performed to improve the resolution, while for the low-risk area with environmental risk entropy lower than the preset environmental risk entropy, the grid is kept or merged to reduce the resolution. Finally, a multi-resolution three-dimensional grid with appropriate density and matching risk distribution is formed, realizing the optimized allocation of computing resources, so that the digital twin model can ensure high-precision simulation of high-risk areas while effectively controlling the geometric complexity of the overall model, effectively improving the rendering efficiency and interactive real-time performance of the subsequent training simulation process.

[0032] Step S15, determining the environmental risk entropy of each grid cell in the multi-resolution three-dimensional grid, loading the preset texture-physical behavior model on the grid cell with environmental risk entropy higher than the preset environmental risk entropy, and loading the preset texture-simplified physical model on the grid cell with environmental risk entropy lower than the preset environmental risk entropy, to obtain the digital twin model.

[0033] It should be noted that the preset environmental risk entropy is a pre-set threshold value for distinguishing high-risk and low-risk levels.

[0034] By querying the environment risk topology map generated in step S13, the environment risk entropy corresponding to each grid cell center point or coverage area in the multi-resolution three-dimensional grid is directly obtained, so as to determine the environment risk entropy of each grid cell.

[0035] During the model loading process, the environment risk entropy of each grid cell is compared with the preset environment risk entropy. For the grid cells with environment risk entropy higher than the preset environment risk entropy, high-precision texture maps are called from the model library and complex physical behavior models are bound, i.e., the preset texture-physical behavior model (such as simulating detailed interactions such as device collision and fluid diffusion), to realistically reproduce the details of high-risk environments. For the grid cells with environment risk entropy lower than the preset environment risk entropy, simplified textures are loaded and basic physical models are associated, i.e., the preset texture-simplified physical model (such as only containing simple collision detection), to quickly render low-risk areas, realize adaptive construction of the digital twin model, and effectively optimize the resource occupation of the overall model while ensuring high simulation and interaction authenticity in high-risk areas, significantly improving the running efficiency and real-time interaction experience of subsequent work injury prevention training simulation.

[0036] Further, step S12 can include steps S121-S123: In step S121, each edge computing node respectively normalizes the corresponding environment data stream to generate a plurality of environment parameter vectors.

[0037] Using the preset parameter range (such as the historical minimum and maximum values of temperature, humidity, light, and gas concentration), the minimum-maximum normalization algorithm is used to linearly transform the value of each environment parameter in the environment data stream to the interval [0, 1], for example, for the temperature parameter, the normalized value calculation formula is (current temperature value-historical minimum temperature) / (historical maximum temperature-historical minimum temperature), thereby eliminating the dimensional differences of different environment parameters, i.e., the normalization operation in this embodiment is to divide the difference between the current value and the historical minimum value by the difference between the historical maximum value and the historical minimum value.

[0038] After processing, each edge computing node combines the normalized environment data stream values at the same time into a standardized vector, i.e., an environment parameter vector, so that environment data of different units and magnitudes have comparability and consistency, providing a standardized data basis for subsequent calculation of environment fluctuation characteristics and environment risk entropy.

[0039] In step S122, the fluctuation amplitudes of each environment parameter vector in each environment dimension within the sliding time window are calculated to obtain a plurality of environment fluctuation characteristics.

[0040] By setting a fixed length of sliding time window, the time series data of each environmental parameter vector in the sliding time window is sequentially intercepted, and the variance of the time series data of each environmental dimension (such as temperature, humidity, light, and gas concentration) in the sliding time window is calculated as a quantitative indicator of the fluctuation amplitude of the corresponding environmental dimension, thereby obtaining the environmental fluctuation characteristics corresponding to each environmental dimension. The normalized environmental parameter vector is converted into a feature sequence that can dynamically reflect the fluctuation intensity of each environmental parameter in the near future, thereby capturing the instability and change trend of the environmental state, and providing a key dynamic input basis for calculating the uncertainty measure representing the comprehensive risk in the subsequent steps.

[0041] In step S123, according to the environmental fluctuation characteristics, the uncertainty measures of each edge computing node in each environmental dimension are calculated, and the uncertainty measures in each environmental dimension are fused to generate the environmental risk entropy corresponding to each edge computing node.

[0042] The environmental fluctuation characteristics calculated in the sliding time window for each environmental dimension are used as the basic data of the state uncertainty of the corresponding environmental dimension, and the information entropy calculation formula is applied to quantify it. For example, for the temperature dimension, its uncertainty measure is calculated by the following formula 1: Formula 1 Wherein, represents the probability of the normalized temperature fluctuation amplitude appearing in the discretization interval, and the larger the temperature fluctuation amplitude, the more unpredictable the temperature fluctuation. represents the uncertainty measure of the temperature dimension. The uncertainty measures of other dimensions are also calculated by the formula shown in formula 1. By substituting the normalized humidity fluctuation amplitude, temperature fluctuation amplitude, and gas concentration amplitude into formula 1, the corresponding uncertainty measure can be calculated.

[0043] After obtaining the uncertainty measures of temperature, humidity, light, and gas concentration of four environmental dimensions, these uncertainty measures are fused by using the weighted summation method, that is, different weight coefficients are preset according to the influence degree of each environmental dimension on the injury risk, and the weighted sum is finally calculated as the environmental risk entropy of the corresponding edge computing node, and the calculation formula is: environmental risk entropy wherein w1, w2, w3, and w4 are the weight coefficients of the four environmental temperatures; represents the uncertainty measure of the humidity dimension; represents the uncertainty measure of the light dimension; To measure the uncertainty on the gas concentration, the uncertainty of multiple environmental dimensions is integrated into a single and quantitative environmental risk entropy, which not only reflects the overall instability degree of the local environment, but also embodies the difference in risk contribution of different environmental factors.

[0044] In a feasible implementation, with reference to Figure 3 The specific execution steps of the simulation module include steps S21-S24 as shown in the figure: Step S21, the digital twin model presents the virtual scene to the user.

[0045] The specific implementation of presenting the virtual scene to the user by the digital twin model is to load the multi-resolution three-dimensional grid and the recorded texture and physical behavior model generated by the construction module through a terminal device such as a computer display, and to calculate and draw a three-dimensional visual scene consistent with the real work site in real time using a graphics rendering technology, thereby converting the digital twin model into a virtual scene that the user can perceive, and realizing an immersive and highly realistic interactive training platform for the user, enabling the user to intuitively familiarize with the layout, equipment and potential hazards of the work site without physical risk, thereby laying a safe and effective visual and interactive foundation for subsequent simulation operations.

[0046] Step S22, based on the real-time operation of the user in the virtual scene of the site, the digital twin model is simulated to obtain the environmental state changes and equipment interaction feedback caused by the real-time operation.

[0047] By continuously monitoring the input instructions of the user on the terminal device and mapping these input instructions into the operation actions of the corresponding objects in the virtual scene, the digital twin model then calculates the operation actions according to its built-in physical rules and logical states, for example, when the user operates the virtual valve, the model will calculate the flow and pressure changes of the medium in the pipeline in real time according to the fluid mechanics equation and update the display state of the related equipment, or when the user approaches a high-risk area, the model will trigger the corresponding sound and light alarm feedback according to the behavior model loaded by the grid unit, creating a dynamic simulation environment that can accurately respond to user behavior in real time, enabling the user's operation to immediately trigger visual environmental state changes and equipment interaction feedback, thereby providing the user with an immersive interactive experience of verifying operation consequences and perceiving the causal relationship between operation and risk.

[0048] Step S23, collect the decision actions of the user under the environmental state changes and equipment interaction feedback.

[0049] After the digital twin model triggers the environmental state update and device feedback, the subsequent operation instruction sequence of the user on the terminal device, such as keyboard operation and mouse click trajectory, is recorded, and the physiological response data obtained through the wearable device, such as eye movement trajectory and heart rate variability, is synchronously collected, so as to completely capture the cognitive judgment and operation response of the user when facing the dynamic simulation scenario. By obtaining the multi-modal data sequence reflecting the real decision logic and stress behavior of the user, an original basis for representing the behavior characteristics of the individual is provided for subsequent quantitative operation analysis, so as to establish a complete data closed loop from the virtual scene feedback to the user behavior response.

[0050] Step S24, quantizing the decision action to obtain behavior data.

[0051] The decision action sequence collected in step S23 is segmented according to a preset time window, the action type in each time window is one-hot encoded, the action duration is quantized at the millisecond level, the operation trajectory coordinates are sequentially recorded, and these numerical values are aligned and integrated into a structured data table according to the time stamp, so as to convert the continuous decision action into discrete and numerical behavior data, and realize the conversion of the unstructured operation behavior of the user into a machine-readable data format.

[0052] In a feasible implementation manner, referring to FIG. 3, Figure 4 The specific execution steps of the analysis module include steps S31-S34 as shown in the figure. Step S31, feature extraction is performed on the behavior data to obtain a behavior feature vector.

[0053] By using a multi-modal feature fusion method, representative features are extracted from the operation instruction sequence and the physiological response data of the behavior data, and then fused. Specifically, the unit time keystroke frequency and combination mode of the keyboard operation sequence are calculated by using a sliding window, the curvature change of the moving trajectory and the click distribution heat map of the mouse click trajectory are calculated, the gaze point distribution entropy and saccade speed variance of the eye movement trajectory are extracted, the time domain standard deviation and the high-low power ratio in the frequency domain of the heart rate variability are calculated, and finally the feature vectors (such as keystroke frequency and mouse trajectory curvature) of the operation instruction sequence and the feature vectors (such as eye movement entropy and heart rate variability index) of the physiological response data are directly spliced to combine a unified behavior feature vector, so as to fuse the discrete behavior data into a comprehensive feature representation that can fully reflect the operation habits and physiological state of the user.

[0054] Step S32, calculating the behavior feature vector by using a preset safety evaluation model to obtain a safety deviation.

[0055] It should be noted that the preset safety evaluation model is a machine learning model pre-trained based on a large amount of standard safety behavior data, and the preset safety evaluation model establishes a decision boundary between normal operation and risk operation by learning the distribution mode of the safety behavior feature vector.

[0056] During calculation, the behavior feature vector is input into the preset safety evaluation model, the preset safety evaluation model processes the behavior feature vector through forward propagation or kernel function transformation, and outputs a continuous value, which represents the deviation degree of the behavior feature vector of the current user from the standard safety behavior mode in the feature space, i.e., the safety deviation degree. In this way, the complex behavior features of the user are converted into a unified and quantifiable safety index, which objectively reflects the difference degree between the user operation and the safety specification.

[0057] In step S33, the safety deviation degree is input into the preset risk evaluation rule for risk matching, and the risk type and the injury risk grade are obtained.

[0058] It should be noted that the preset risk evaluation rule is an if-then rule base constructed based on domain expert knowledge, which clearly defines the injury risk grade division standards corresponding to different safety deviation degree intervals, such as dividing the safety deviation degree 0-0.2 into low injury risk, 0.2-0.5 into medium injury risk, and 0.5 or more into high injury risk, and classifying the risk types corresponding to various operation behavior modes, such as operation sequence error and response delay too long.

[0059] During risk matching, the safety deviation degree is compared with the preset threshold interval in the preset risk evaluation rule base to determine the injury risk grade to which it belongs, and the abnormal combination mode of a specific feature in the behavior feature vector, such as high-frequency operation combined with abnormal eye movement trajectory, is matched with the risk feature mode defined in the preset risk evaluation rule base to determine the specific risk type, thereby converting the quantified safety deviation degree into a risk evaluation conclusion with clear semantic interpretation, so that the system can not only judge the risk size but also identify the risk nature, providing classification and grading decision basis for generating targeted injury warning information.

[0060] In step S34, the injury warning information is generated according to the risk type and the injury risk grade.

[0061] By querying the preset warning information template library, the preset warning information template library predefines the corresponding risk description text, severity identifier and improvement measure suggestion for each risk type and injury risk grade combination.

[0062] According to the risk type and the work injury risk level, a corresponding template is matched from a preset early warning information template library, and specific parameters such as risk position and deviation value are automatically filled in to generate a structured work injury early warning information. The work injury early warning information usually includes risk classification, risk level, specific performance and recommended action, etc., so as to realize the conversion of abstract risk assessment results into personalized early warning feedback that can be intuitively understood by users and has operation guidance significance.

[0063] In a feasible implementation, referring to Figure 5 As shown in the figure, the specific execution steps of the output module include steps S41-S44: Step S41, data quantization processing is performed on the work injury early warning information and the historical training record to obtain safety evaluation data.

[0064] The purpose of data quantization processing on the work injury early warning information and the historical training record is to convert discrete safety information of different types and sources into a unified numerical expression form, thereby providing a data basis for subsequent safety credit score calculation.

[0065] Step S42, based on the safety evaluation data, weight distribution is performed on the work injury early warning information and the historical training record to obtain a weighted evaluation value.

[0066] According to the time attribute and risk severity of the safety evaluation data, a weight distribution rule is established. First, a time decay function is applied to the historical training record, so that the recent training performance obtains a higher weight. At the same time, a differentiated weight coefficient is set according to the risk level of the work injury early warning information, and a higher weight corresponds to a high-risk early warning. Then, each type of safety evaluation data is multiplied by its corresponding weight, and finally a weighted sum is calculated to obtain a weighted evaluation value that comprehensively reflects the user's recent safety performance, so that the recent performance and serious risk behavior have a greater impact on the score result, thereby ensuring that the safety credit score can timely reflect the changes in the user's current safety status.

[0067] Step S43, according to the weighted evaluation value, the score of the user in the user group to which the user belongs is calculated to obtain a relative safety score.

[0068] First, the weighted evaluation values of all users in the user group to which the user belongs, such as the same post or the same department, are collected. Then, the percentile ranking algorithm is used to compare the weighted evaluation value of the current user with the weighted evaluation values of other users in the group, and the specific percentile ranking of the user score in the group is calculated. For example, if the weighted evaluation value of the user exceeds that of 80% of the users in the group, the relative safety score of the user is recorded as 80. In this way, the absolute safety performance of the user is converted into a relative position ranking in a specific reference group, effectively eliminating the influence of fluctuations in the overall level of the group or differences in risk of different posts on the evaluation result, thereby ensuring that the final safety credit score can reflect individual performance.

[0069] Step S44, mapping the relative safety score to a preset score range to obtain a safety credit score.

[0070] The relative safety score in percentage is proportionally mapped to the preset score range by a linear conversion function, for example, when the preset score range is 0-100, a direct mapping is performed by the formula "safety credit score = relative safety score", or when the preset score range is 0-1000, a linear amplification is performed by "safety credit score = relative safety score x 10", to ensure that the score and the credit maintain a strict linear correspondence.

[0071] Further, step S41 can include steps S411-S413: Step S411, by a preset quantification index system, the warning times, warning levels, training scores and training completion degrees are uniformly dimensionally processed to obtain a quantification index set.

[0072] The warning times are converted to a value in the range of 0-1 by a linear normalization method, different warning levels (such as high, medium and low) are mapped to corresponding numerical coefficients (such as 1.0, 0.6 and 0.2), the training scores in percentage are converted to standard scores, and the training completion degrees are directly quantified according to the completion percentage.

[0073] All the indexes are uniformly converted to comparable dimensionless values by the above regularization processing to form the quantification index set, so as to eliminate the dimensional differences and order of magnitude differences between different sources of indexes, and make the work injury warning information and historical training records these multi-source heterogeneous data be able to be comprehensively operated under the same standard, to provide a fair and reliable numerical basis for subsequent weighted fusion.

[0074] Step S412, weighted fusion is performed on each quantification index in the quantification index set to generate a comprehensive safety evaluation value.

[0075] By a preset weight distribution scheme, each quantification index in the quantification index set, i.e. the warning times, the warning levels, the training scores and the training completion degrees, is respectively assigned a corresponding weight coefficient, and these weight coefficients are set based on the importance of each quantification index to the overall safety evaluation. Then, a weighted average algorithm is used to multiply the value of each quantification index by the corresponding weight coefficient, sum up the products, and divide the sum by the total of the weight coefficients to obtain the comprehensive safety evaluation value, so as to integrate multiple independent quantification indexes into a unified comprehensive score according to their importance, thereby comprehensively and reasonably reflecting the overall safety status of the user.

[0076] Step S413, mapping the comprehensive safety evaluation value to a preset safety evaluation range to obtain a safety evaluation data.

[0077] The comprehensive safety evaluation value is scaled according to the minimum value and the maximum value of the preset safety evaluation range through a linear conversion function, for example, if the original range of the comprehensive safety evaluation value is [0, 100] and the preset safety evaluation range is [0, 10], the formula "safety evaluation data = (comprehensive safety evaluation value / 100) x 10" is used for calculation, so as to map the comprehensive safety evaluation value to the preset safety evaluation range, standardize the comprehensive safety evaluation value to a unified numerical interval, eliminate the scale difference between different evaluation values, and ensure that all data have consistent comparability and operability.

[0078] The above are preferred embodiments of the present application, and do not limit the protection scope of the present application, so: all equivalent changes made according to the structure, shape, principle of the present application should be covered within the protection scope of the present application.

Claims

1. A workplace injury prevention training system, characterized in that, include: A construction module is used to acquire environmental data streams from the actual work site and construct a digital twin model corresponding to the actual work site based on the environmental data streams. The simulation module is used to conduct work injury prevention training for users through the digital twin model, and to collect user behavior data during the work injury prevention training process; The analysis module is used to perform work injury analysis based on the behavioral data and generate work injury early warning information; The output module is used to generate a safety credit score based on the work injury early warning information and historical training records.

2. The work injury prevention training system according to claim 1, characterized in that, The steps of acquiring environmental data streams from the actual workplace and constructing a digital twin model corresponding to the actual workplace based on the environmental data streams include: By acquiring the environmental data stream at each of the sensor network points deployed on the actual work site; Based on the edge computing nodes embedded in each of the sensor network points, the environmental data streams are processed to generate multiple environmental risk entropies. Based on the multiple environmental risk entropies, an environmental risk topology map of the actual work site is generated; Based on the environmental risk entropy in the environmental risk topology map, a grid resolution division threshold is set to re-divide the preset three-dimensional network of the real work site to obtain a multi-resolution three-dimensional grid. The environmental risk entropy of each grid cell in the multi-resolution 3D mesh is determined. A preset texture-physical behavior model is loaded on the grid cells whose environmental risk entropy is higher than the preset environmental risk entropy, and a preset texture-simplified physical model is loaded on the grid cells whose environmental risk entropy is lower than the preset environmental risk entropy, to obtain the digital twin model.

3. The work injury prevention training system according to claim 2, characterized in that, The step of processing each environmental data stream based on edge computing nodes embedded in each of the sensor network points to generate multiple environmental risk entropies includes: Each edge computing node performs normalization processing on the corresponding environmental data stream to generate multiple environmental parameter vectors. Calculate the fluctuation amplitude of each environmental parameter vector in each environmental dimension within the sliding time window to obtain multiple environmental fluctuation characteristics; Based on the environmental fluctuation characteristics, the uncertainty measure of each edge computing node in each environmental dimension is calculated, and the uncertainty measures in each environmental dimension are fused to generate the environmental risk entropy corresponding to each edge computing node.

4. The work injury prevention training system according to claim 1, characterized in that, The steps of conducting workplace injury prevention training for users through the digital twin model and collecting user behavioral data during the workplace injury prevention training process include: The digital twin model is used to present a virtual scene of the venue to the user; Based on the user's real-time operations in the virtual scene of the venue, the digital twin model is used for simulation to obtain the environmental state changes and device interaction feedback caused by the real-time operations. Collect user decision-making actions in response to changes in the environmental state and feedback from device interactions; The decision-making action is quantified to obtain the behavioral data.

5. The work injury prevention training system according to claim 1, characterized in that, The steps of performing work injury analysis based on the behavioral data and generating work injury early warning information include: Feature extraction is performed on the behavioral data to obtain a behavioral feature vector; The safety deviation is obtained by calculating the behavioral feature vector using a preset safety assessment model. The safety deviation is input into a preset risk assessment rule for risk matching to obtain the risk type and work injury risk level; The work injury early warning information is generated based on the risk type and the work injury risk level.

6. The work injury prevention training system according to claim 1, characterized in that, The step of generating a safety credit score based on the work injury early warning information and historical training records includes: The work injury early warning information and the historical training records are subjected to data quantification processing to obtain safety assessment data; Based on the safety assessment data, the work injury early warning information and the historical training records are weighted and assigned to obtain a weighted assessment value. Based on the weighted evaluation value, the user's score within their respective user group is calculated to obtain a relative security score; The relative security score is mapped to a preset score range to obtain the security credit score.

7. The work injury prevention training system according to claim 6, characterized in that, The work injury early warning information includes the number of early warnings and the early warning level; the historical training records include training performance and training completion rate; the step of quantifying the work injury early warning information and the historical training records to obtain safety assessment data includes: By using a pre-defined quantitative indicator system, the number of warnings, the warning level, the training results, and the training completion rate are processed with unified dimensions to obtain a set of quantitative indicators. The quantitative indicators in the set of quantitative indicators are weighted and fused to generate a comprehensive security assessment value; The comprehensive security assessment value is mapped to a preset security assessment range to obtain the security assessment data.

Images