Training deduction management method and system based on scene data fusion

Abstract

The application discloses a training deduction management method and system based on scene data fusion, and relates to the technical field of big data processing. The method comprises the following steps: S1, collecting deduction scene data in real time, and performing data preprocessing to obtain a deduction scene data set; S2, identifying corresponding events to be executed for current training deduction, and constructing an event list to be executed; S3, analyzing the completion, process compliance and timeliness of the events to be executed, marking abnormal events, forming an event detection result, and counting the number of abnormal events and completed events; S4, comprehensively evaluating the effect of the current deduction, forming a deduction evaluation record, identifying weak training events and key improvement posts, and generating a closed-loop improvement task. The method solves the problem that the existing training deduction platform only evaluates according to triggered operations, and cannot identify missing or abnormal events to be executed, resulting in inaccurate positioning of weak links in the deduction evaluation result.

Description

Technical Field

[0001] This invention relates to the field of big data processing technology, specifically to a training simulation management method and system based on scenario data fusion. Background Technology

[0002] With the increasing demand for training in emergency management, public safety, fire fighting, and accident emergency response, training simulations based on information platforms have become an important means to improve personnel's emergency response capabilities and collaborative response levels. During training simulations, accident scenario data, trainee operational data, standard handling procedure data, and training assessment data are continuously generated. These data sources are diverse, the volume is large, and the relationships are complex. It is necessary to utilize big data resource services to uniformly collect, store, and access this multi-source data during the training simulation process, and to use data mining software to organize, correlate, classify, statistically analyze, and pattern-following the simulation data. This provides data support for training assessment, capability evaluation, and subsequent training improvement, and is widely used in fields such as fire emergency response, production safety, energy and chemical industry, rail transportation, and public safety.

[0003] For example, Chinese patent CN112287176A discloses a method for building and training a training platform for troubleshooting defects in the secondary system of an intelligent substation. The method includes the following steps: S1, obtaining SCD files from multiple manufacturers and parsing each SCD file until the entire physical link of the substation is visualized; S2, establishing a secondary system defect database based on the parsed SCD files; S3, designing a fault simulation module for the secondary circuit of the intelligent substation based on the secondary system defect database. This simulation module is used for teaching demonstrations and exam simulations. A training method for troubleshooting defects in the secondary system of an intelligent substation is also provided, using the platform built using the above method for training. This invention can cultivate composite talents integrating intelligent equipment and relay protection professional skills, significantly improve the training capabilities of intelligent substation secondary system training bases, increase the efficiency of practical training, reduce training costs, and improve the troubleshooting capabilities of operation and maintenance personnel without affecting equipment lifespan and steady-state operation, thus accelerating the cultivation of highly skilled relay protection personnel.

[0004] For example, Chinese patent CN118069707A discloses an integrated teaching system for military simulation exercises, research, and training, including modules for image interpretation, logistics database, logistics service calculation, logistics document drafting, map drawing, official correspondence, and online examination. This system provides a model and guide for multi-level, three-dimensional military simulation exercises, research, and training. The system adopts a comprehensive and in-depth approach to "deep learning teaching improvement," promoting the integration of teaching and research. In the later stages of the project, the concept of deep systems will be integrated into regular classroom teaching. Each subject will integrate deep learning teaching improvement cases with subject-specific teaching and research courses in a unified design and implementation. The teaching and research themes focus on core competency-oriented deep learning teaching strategies, with military simulation teaching and research content unfolding layer by layer around the elements of the practical exercise model. The teaching and research format embodies collaborative discussion, systematic guidance, and multi-party interaction.

[0005] However, existing simulation platforms drive training through incident chains and record the actions triggered by trainees, using these as the basis for scoring and incident control determination. When trainees only complete some visible actions, such as fire extinguishing, but omit key actions strongly related to the scenario state, such as fire separation, evacuation of key areas, broadcast notifications, and verification receipts, the lack of rule verification and missing detection for "should-be-executed events" makes it impossible to identify gaps where "should-be-executed events are not executed correctly." This could lead to the incident being deemed under control and receiving an excessively high score even when key events are missing, resulting in inexplicable and unreproducible evaluation results, and making it difficult to pinpoint weaknesses in training.

[0006] Therefore, in order to address the above issues, there is an urgent need for training simulation management methods and systems based on scenario data fusion. Summary of the Invention

[0007] Technical problems to be solved

[0008] To address the shortcomings of existing technologies, this invention provides a training simulation management method and system based on scenario data fusion. This solves the problem that existing training simulation platforms only evaluate based on triggered operations, making it difficult to identify missing or abnormal execution events, resulting in distorted simulation evaluation results and inaccurate identification of weak links.

[0009] Technical solution

[0010] To achieve the above objectives, the present invention provides the following technical solution: a training simulation management method based on scenario data fusion, comprising the following steps: S1, real-time collection of simulation scenario data during the training simulation process, and preprocessing of the simulation scenario data to obtain a simulation scenario dataset; S2, based on the simulation scenario dataset, identifying the events to be executed corresponding to the current training simulation, and constructing a list of events to be executed; S3, combining the list of events to be executed with the simulation scenario dataset, analyzing the completion status, process compliance, and timeliness of the events to be executed, marking abnormal events, forming event detection results based on the abnormal event markings, and statistically analyzing the number of abnormal events and completed events; S4, combining the statistical results of the number of abnormal events and completed events with the simulation scenario dataset, comprehensively evaluating the effect of this round of simulation, and forming a simulation evaluation record; combining historical simulation evaluation records, identifying weak training events and key improvement positions, and generating closed-loop improvement tasks.

[0011] Furthermore, the specific process of real-time collection of simulation scenario data during training simulations is as follows: Real-time collection of simulation scenario data specifically includes: reading the current accident type from the currently loaded simulation script and reading the simulation start time from the simulation platform system clock; reading the trainee number, job number, operation event number, operation time, and operation status from the simulation platform interaction log, with the operation status including execution success and execution failure; and reading standard handling process data from the simulation rule base, including standard handling event number, applicable accident type, responsible job number, preceding event number, event standard completion time limit, and event importance level.

[0012] Further, the specific process of preprocessing the simulation scenario data to obtain the simulation scenario dataset is as follows: The simulation scenario data is converted into a unified encoding format and a unified time unit; using the start time of this round of simulation as the unified time zero point, the start time of this round of simulation is subtracted from each operation time to obtain the corresponding cumulative operation time; for duplicate data records with the same student ID, job ID, and operation event ID, only the record with the largest cumulative operation time is retained; when the operation status is empty, the operation status is marked as not returned; when the preceding event ID is empty, the preceding event ID is marked as no preceding event; based on the preprocessed simulation scenario data, a simulation scenario dataset is formed and written into the training simulation management database.

[0013] Furthermore, based on the simulation scenario dataset, the specific process for identifying the events to be executed corresponding to the current training simulation and constructing a list of events to be executed is as follows: The current accident type is compared with the applicable accident types in the standard handling process data, and the current job number is compared with the responsible job number in the standard handling process data; when the applicable accident type is the same as the current accident type, and the responsible job number is the same as the current job number, the corresponding standard handling event number is marked as the current event to be executed; the preceding event number corresponding to the event to be executed is read, and the standard handling process sequence is formed based on the correlation between the preceding event numbers, serving as the basis for identifying subsequent events that exceed the specified level; the responsible job number, preceding event number, event standard completion deadline, and event importance level corresponding to each event to be executed are read, associated, and stored to form a list of events to be executed, which is then written into the training simulation management database.

[0014] Furthermore, combining the list of events to be executed and the simulation scenario dataset, the completion status, process compliance, and timeliness of the events to be executed are analyzed. The specific process for marking abnormal events is as follows: For each event to be executed, abnormal events are marked: The standard handling event number corresponding to the event to be executed is compared with the operation event number in the simulation scenario dataset; when there is no operation event number in the simulation scenario dataset that is the same as the standard handling event number, the current event to be executed is marked as an unexecuted event; when there is an operation event number in the simulation scenario dataset that is the same as the standard handling event number, the operation status corresponding to the operation event number is read; when the operation status is not successful, the current event to be executed is marked as an unsuccessful execution event; the preceding event number corresponding to each event to be executed is read; when the preceding event number is marked as having a preceding event, it is determined whether there is an operation event number in the simulation scenario dataset that is the same as the preceding event number, and whether the corresponding operation status is successful; if the conditions are not met and the operation status of the operation event number corresponding to the current event to be executed is successful, the current event to be executed is marked as a skipped execution event; at the same time, for each event to be executed, timeout execution events are identified.

[0015] Furthermore, for each event to be executed, the specific process for identifying timeout execution events is as follows: For each event to be executed, when there is an operation event number in the simulation scenario dataset that is the same as the standard handling event number, and the operation status is successful, the cumulative operation time corresponding to the successfully executed operation event number is read. The cumulative operation time corresponding to the current operation event number is subtracted from the cumulative operation time corresponding to the previous operation event number to obtain the event execution time. The previous operation event number is determined according to the order of cumulative operation times in the simulation platform interaction log. If there is no previous operation event number, the corresponding cumulative operation time is used as the event execution time. The event execution time is compared with the standard completion time limit corresponding to the current event to be executed. When the event execution time is greater than the standard completion time limit, the current event to be executed is marked as a timeout execution event. For timeout execution events, the event execution time is subtracted from the standard completion time limit to obtain the event timeout duration.

[0016] Furthermore, the specific process for generating event detection results based on abnormal event tags and counting the number of abnormal and completed events is as follows: When an event to be executed does not have any abnormal event tags, it is determined to be a completed event; when the same event to be executed simultaneously meets two or more abnormal event tags, all corresponding abnormal event tags are retained; the current list of events to be executed is read and associated with the corresponding abnormal event tags and event timeout duration to form event detection results; the event detection results are classified according to abnormal event tags, and events to be executed marked as unexecuted events, unsuccessful execution events, timed-out execution events, and out-of-level execution events are written into the critical event abnormality list, and events to be executed judged as completed events are written into the completed event list; the number of unexecuted events, unsuccessful execution events, timed-out execution events, out-of-level execution events, and completed events are counted respectively, and the number of unexecuted events, unsuccessful execution events, timed-out execution events, and out-of-level execution events are added together to obtain the number of abnormal events; the event detection results, the critical event abnormality list, and the completed event list are written into the training simulation management database.

[0017] Furthermore, combining the statistical results of the number of abnormal events and completed events with the simulation scenario dataset, a comprehensive evaluation of the simulation effect is conducted, and the specific process for forming the simulation evaluation record is as follows: The number of unexecuted events, unsuccessful execution events, timed-out execution events, and out-of-level execution events in the critical event anomaly list, as well as the event importance level corresponding to each abnormal event, are read; simultaneously, the number of completed events and the event importance level corresponding to each completed event are read from the completed event list; the event importance levels of all currently executable events are summed to obtain the total importance level of executable events; the event importance levels of all completed events are summed to obtain the total importance level of completed events; and the event importance levels of all timed-out execution events and out-of-level execution events are recalculated. To obtain the total importance level of process abnormal events, sum the importance levels of the events. Divide the total importance level of completed events by the total importance level of events to be executed, and multiply by 100 to obtain the simulation completion score. Divide the total importance level of process abnormal events by the total importance level of events to be executed, and multiply by 100 to obtain the process abnormality deduction value. Subtract the process abnormality deduction value from 100 to obtain the process standardization score. When the process standardization score is less than zero, correct it to zero. Link and store the current incident type, trainee number, job number, standard handling event number corresponding to the event to be executed, critical event abnormality list, completed event list, simulation completion score, and process standardization score to form a simulation evaluation record, and write it into the training simulation management database.

[0018] Furthermore, based on historical simulation and evaluation records, the specific process for identifying weak training events and key improvement positions, and generating closed-loop improvement tasks is as follows: Based on historical simulation and evaluation records, the list of historical critical event anomalies is categorized and statistically analyzed according to the standard handling event number, and the number of historical anomaly markers corresponding to each standard handling event is calculated; when the number of anomaly markers corresponding to the same standard handling event exceeds the event anomaly threshold, the corresponding standard handling event number is marked as a weak training event; the list of historical critical event anomalies is categorized and statistically analyzed according to the responsibility position number, and the number of anomalies corresponding to each responsibility position is calculated separately; when the number of anomalies corresponding to the same responsibility position exceeds the anomaly number threshold, the corresponding responsibility position is marked as a key improvement position; for key improvement positions and weak training events, a closed-loop improvement task list is generated and output.

[0019] The second aspect of this invention provides a training simulation management system based on scenario data fusion, comprising: a simulation data acquisition and processing module, used to acquire simulation scenario data in real time during the training simulation process, and preprocess the simulation scenario data to obtain a simulation scenario dataset; an execution event identification module, used to identify the execution events corresponding to the current training simulation based on the simulation scenario dataset, and construct an execution event list; a key event anomaly detection module, used to combine the execution event list and the simulation scenario dataset to analyze the completion status, process compliance, and timeliness of the execution events, mark anomalies, form event detection results based on the anomaly event markings, and perform statistical analysis of the number of anomalies and completed events; and a simulation evaluation closed-loop management module, used to combine the statistical results of the number of anomalies and completed events with the simulation scenario dataset to comprehensively evaluate the effect of this round of simulation, and form a simulation evaluation record; and to combine historical simulation evaluation records to identify weak training events and key improvement positions, and generate closed-loop improvement tasks.

[0020] Beneficial effects

[0021] The present invention has the following beneficial effects:

[0022] (1) This invention forms a list of events to be executed by matching the current accident type with the standard handling process, so that the system no longer evaluates based solely on the operations already triggered by the trainee, but can cover key handling events that should be completed in the standard process but were not actively triggered by the trainee, thereby reducing the omission of key events.

[0023] (2) By comparing the event to be executed with the operation event number, operation status, cumulative operation time and preceding event number, this invention can distinguish different abnormal types such as non-execution, unsuccessful execution, timeout execution and over-level execution, avoiding the use of "whether there is an operation record" as the sole criterion for judgment, and improving the interpretability of event detection results.

[0024] (3) The present invention combines completed events, abnormal events and event importance levels to form a simulation completion score and a process standardization score, which transforms the simulation evaluation from subjective scoring to a structured evaluation based on event status and process abnormalities. This helps to reduce scoring bias and ensures that the same simulation record yields consistent results when evaluated repeatedly.

[0025] (4) This invention classifies and statistically analyzes abnormal events in historical simulation evaluation records, identifies weak training events and key improvement positions, and generates closed-loop improvement tasks. This enables the system to not only evaluate the results of a single simulation, but also to continuously locate high-frequency weak links, providing a basis for subsequent targeted training and simulation scheme optimization.

[0026] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0027] Figure 1 A flowchart for a training simulation management method based on scenario data fusion;

[0028] Figure 2 A structural diagram of a training simulation management system based on scenario data fusion;

[0029] Figure 3 This is a diagram illustrating the standard handling procedure and the results of critical event anomaly detection.

[0030] Figure 4 This is a schematic diagram of the event-driven inference and evaluation process. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. As those skilled in the art will understand, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] Please see Figures 1-4 This invention provides a technical solution: a training simulation management method based on scenario data fusion, such as... Figure 1 As shown, the process includes the following steps: S1, real-time collection of simulation scenario data during the training simulation process, and preprocessing of the simulation scenario data to obtain a simulation scenario dataset; S2, based on the simulation scenario dataset, identifying the events to be executed corresponding to the current training simulation, and constructing a list of events to be executed; S3, combining the list of events to be executed and the simulation scenario dataset, analyzing the completion status, process compliance, and timeliness of the events to be executed, marking abnormal events, forming event detection results based on the abnormal event markings, and statistically analyzing the number of abnormal events and completed events; S4, combining the statistical results of the number of abnormal events and completed events with the simulation scenario dataset, comprehensively evaluating the effect of this round of simulation, forming a simulation evaluation record; combining historical simulation evaluation records, identifying weak training events and key improvement positions, and generating closed-loop improvement tasks.

[0033] Specifically, the process of real-time collection of simulation scenario data during training simulations is as follows: Real-time collection of simulation scenario data includes: reading the current accident type from the currently loaded simulation script. The currently loaded simulation script is the accident scenario file called by the simulation platform before the start of this round of training simulations, and the current accident type is the accident category identifier pre-configured in the accident scenario file. Taking a fire emergency scenario as an example, the current accident type is the accident category identifier pre-configured in the fire emergency scenario file, such as initial fires in commercial complexes, vehicle fires in underground parking garages, goods fires in warehouse areas, electrical equipment short-circuit fires, or kitchen fume duct fires, etc., which are fire emergency training accident types. The simulation start time is also read from the simulation platform's system clock. The simulation start time is the system time when this round of training simulations is started and the first system operation record is generated, used as the time benchmark for unified conversion of subsequent operation times. The simulation platform's interaction log is used to retrieve trainee ID, job ID, operation event ID, operation time, and operation status. This log is automatically generated when trainees perform actions such as clicking, confirming, reporting, controlling, and issuing commands on the simulation interface. The trainee ID identifies the participant performing the action, the job ID identifies their role in the simulation, the operation event ID identifies the standard handling event corresponding to the action, the operation time records the system time when the action is received by the platform, and the operation status records the result, including at least success and failure. Standard handling process data is retrieved from the simulation rule base, including the standard handling event ID, applicable accident type, responsible job ID, preceding event ID, event standard completion deadline, and event importance level. The simulation rule base is a pre-configured set of standard handling procedures. This rule base is pre-established based on the emergency plans, job handling procedures, training assessment standards, and simulation script configuration rules adopted by the training unit, and is loaded by the simulation platform before the training simulation begins. The standard handling event number uniquely identifies each handling event in the standard handling procedure; the applicable accident type limits the accident category to which the corresponding standard handling event applies; the responsible position number limits the executing position corresponding to the standard handling event; the preceding event number indicates the previous handling event that needs to be completed before the standard handling event is executed; the event standard completion time limit indicates the maximum allowed time for the standard handling event to complete after the corresponding preceding event is completed, preferably in seconds; and the event importance level indicates the degree of influence of the standard handling event on the simulation evaluation results, preferably set to level 1 to 5, where a higher value indicates a higher degree of influence of the corresponding standard handling event on the accident handling results and simulation evaluation results. Both the event standard completion time limit and the event importance level are preset by the simulation rule base.

[0034] This implementation plan, by clearly defining the data sources of the simulation script, system clock, interaction logs, and simulation rule base, can reliably acquire data on the current accident type, trainee operation records, and standard handling procedures. Furthermore, by establishing traceable data associations through trainee ID, job ID, operation event ID, preceding event ID, and event standard completion time limit, it improves the completeness, standardization, and feasibility of simulation scenario data collection, providing an accurate and unified data foundation for subsequent event identification, abnormal event detection, and simulation evaluation.

[0035] Specifically, the preprocessing of the simulation scenario data to obtain the simulation scenario dataset involves converting the simulation scenario data into a unified encoding format and a unified time unit. The unified encoding format ensures that the current accident type, trainee number, job number, operation event number, operation status, standard handling event number, applicable accident type, responsible job number, preceding event number, and event importance level are represented using the simulation platform's preset encoding rules. UTF-8 encoding is preferred to ensure consistency of Chinese event names, job names, and status names when transmitted between different system modules. The preferred unified time unit is seconds. Using the start time of this simulation round as the unified time zero point, each operation time is subtracted from the start time of this simulation round to obtain the corresponding cumulative operation time. The cumulative operation time represents the occurrence time of the corresponding operation event relative to the start time of this simulation round, facilitating a unified comparison of the time sequence of different operation events. For duplicate data records with the same trainee ID, job title, and operation event number, only the record with the longest cumulative operation time is retained. This record represents the last operation result of the same trainee for the same operation event within the same job title. Since the same trainee may submit multiple times, re-operate after failure, refresh status, or confirm multiple times for the same operation event within the same job title, the simulation platform's interaction log may contain multiple data records with the same trainee ID, job title, and operation event number. The record with the longest cumulative operation time represents the final execution result of the operation event in this round of training simulation, preventing the same operation event from being repeatedly included in subsequent event identification, abnormal event marking, and simulation evaluation processes. Since the subsequent completion status judgment, execution failure judgment, and timeout judgment of this invention are all based on the final state of the operation event, retaining the record with the longest cumulative operation time reflects the final handling result of the corresponding operation event. When the operation status is empty, it is marked as "not returned," indicating that the simulation platform did not receive a clear success or failure result after the corresponding operation event occurred. When the preceding event number is empty, it is marked as "no preceding event," otherwise it is marked as "a preceding event exists." "No preceding event" indicates that the corresponding standard handling event does not depend on the completion result of other standard handling events and can be used as the starting handling event in the corresponding standard handling process. Based on the preprocessed simulation scenario data, a simulation scenario dataset is formed and written into the training simulation management database. The simulation scenario dataset includes the current accident type, trainee number, job number, operation event number, cumulative operation time, operation status, and standard handling process data after unifying coding, unifying time, handling duplicate records, and marking null values. This data is used for subsequent event identification, critical event anomaly detection, and simulation evaluation.

[0036] In this implementation plan, by using unified coding, unified time base, duplicate record processing, and null value marking, the simulation scenario data has a consistent data format and a comparable time order. This avoids interference from repeated operations, inconsistent time formats, and missing states that could lead to subsequent judgments, and improves the standardization, accuracy, and traceability of the simulation scenario dataset. It provides a reliable data foundation for the identification of events to be executed, the detection of anomalies in key events, and the simulation evaluation.

[0037] Specifically, based on the simulation scenario dataset, the process of identifying the events to be executed corresponding to the current training simulation and constructing a list of events to be executed is as follows: The current accident type is compared with the applicable accident types in the standard handling procedure data, and the current job number is compared with the responsible job number in the standard handling procedure data. The current job number is obtained from the job number in the simulation scenario dataset and represents the role of the currently logged-in trainee in this round of training simulation. The applicable accident type limits the types of simulation accidents to which the standard handling event applies, and the responsible job number limits the execution position corresponding to the standard handling event. When the applicable accident type is the same as the current accident type, and the responsible job number is the same as the current job number, the corresponding standard handling event number is marked as the current event to be executed. The current event to be executed represents the standard handling event that the trainee in this position needs to complete and undergo verification in this round of training simulation, under the combined constraints of the current accident type and the current job number. The preceding event number corresponding to the event to be executed is read, and a standard handling process sequence is formed based on the relationship between the preceding event numbers, which serves as the basis for identifying subsequent out-of-level execution events. The relationship between the preceding event numbers is used to represent the sequential dependency between different standard handling events. When the preceding event number of an event to be executed has no preceding event, the corresponding event to be executed is used as the starting event in the corresponding standard handling process sequence. When the preceding event number of an event to be executed does not have no preceding event, the standard handling event corresponding to the preceding event number is used as the previous process event of the current event to be executed. In the standard handling process data, the preceding event numbers are recorded in the form of directed associations. Each directed association points from the preceding event number to the corresponding standard handling event number, indicating the execution order between standard handling events. When a standard handling event is configured with only one preceding event number, the standard handling event corresponding to that preceding event number is considered the unique previous process event. When a standard handling event is configured with two or more preceding event numbers, the two or more preceding event numbers are combined into a preceding event set, and the default rule of "all must be satisfied" is adopted, meaning that the current standard handling event is considered to meet the process order requirements only after all preceding events in the preceding event set have been successfully executed. When optional preceding events need to be configured, an optional preceding event flag is set for the corresponding standard handling event in the inference rule base, and any satisfying rule is adopted, meaning that the current standard handling event is considered to meet the process order requirements only after any preceding event in the preceding event set has been successfully executed. For standard handling processes with branch processes, the preceding event number corresponding to the branch trigger condition is used as the branch starting point, and the standard handling process order of each branch is formed according to the directed associations between the preceding event numbers and the standard handling event numbers.The system reads and associates the responsible position number, preceding event number, standard completion deadline, and importance level of each event to be executed, forming a current event to be executed list, which is then written into the training simulation management database. This current event to be executed list records the standard handling events that trainees in the current position should verify in this round of training simulations, along with their executing positions, process dependencies, deadlines, and importance, providing a data foundation for subsequent monitoring of completion status, process compliance, and timeliness.

[0038] In this implementation plan, by matching the current accident type with the current job number, the standard handling events that trainees in the current job should complete in this round of simulation can be accurately identified, avoiding the misinclusion of other job tasks in the evaluation scope. At the same time, the standard handling process sequence is formed based on the preceding event number, providing a clear basis for subsequent judgments of overstepping the level of execution, exceeding the time limit, and missing events, thereby improving the accuracy of identifying events that should be executed and the pertinence of job-specific assessments.

[0039] Specifically, combining the list of events to be executed and the simulation scenario dataset, the completion status, process compliance, and timeliness of the events to be executed are analyzed. The specific process for marking abnormal events is as follows: For each event to be executed, an abnormal event is marked: the standard handling event number corresponding to the event to be executed is compared with the operation event number in the simulation scenario dataset; the standard handling event number comes from the current list of events to be executed, and the operation event number comes from the interaction log of the simulation platform. Both are represented using a unified encoding format to determine whether the trainee in the current position has triggered the corresponding standard handling event; when there is no operation event number in the simulation scenario dataset that is the same as the standard handling event number, the current event to be executed is marked as an unexecuted event; the unexecuted event indicates that the event to be executed was not triggered by the trainee in the current position during this round of training simulation. When an operation event number with the same standard handling event number exists in the simulation scenario dataset, the operation status corresponding to the operation event number is read. The operation status indicates the execution result after the corresponding operation event is received by the simulation platform. When the operation status is not "execution successful," the current event to be executed is marked as an "execution unsuccessful" event. An "execution unsuccessful" event indicates that the trainee in the current position has triggered the event to be executed, but the simulation platform has not returned a success result. The preceding event number corresponding to each event to be executed is read. When the preceding event number is marked as having a preceding event, it is determined whether an operation event number with the same preceding event number exists in the simulation scenario dataset, and whether the corresponding operation status is "execution successful." The preceding event number indicates the previous process event in the standard handling process sequence for the current event to be executed. This determination is used to confirm whether the previous process event has been completed. If the conditions are not met and the operation status of the operation event number corresponding to the current event to be executed is "execution successful," the current event to be executed is marked as a "skipped execution event." A skipped execution event indicates that the current event to be executed has been executed successfully, but its preceding event has not been executed successfully, indicating that the current operation does not conform to the standard handling process sequence. Simultaneously, for each event that should be executed, a timeout execution event is identified. The timeout execution event indicates that although the current event was successfully executed, its execution time exceeded the corresponding event's standard completion time limit.

[0040] In this implementation plan, by comparing the corresponding execution events with the actual operation events, it is possible to distinguish abnormal situations such as non-execution, unsuccessful execution, execution beyond the level, and execution exceeding the time limit, avoiding simple evaluation based solely on the existence of operation records; at the same time, by combining the preceding event number and the operation status to judge the compliance of the process, the identification of key event anomalies is more accurate and interpretable, providing a reliable basis for subsequent simulation scoring correction, weak link location, and closed-loop improvement.

[0041] Specifically, for each event to be executed, the process for identifying timed-out execution events is as follows: For each event to be executed, when there is an operation event number in the simulation scenario dataset that is the same as the standard handling event number, and the operation status is "executed successfully," it indicates that the trainee in the current position has completed the handling operation corresponding to the event to be executed, and the conditions for timeliness judgment are met; the cumulative operation time corresponding to the successfully executed operation event number is read, and the cumulative operation time is used to represent the occurrence time of the corresponding operation event relative to the start time of this round of simulation; the cumulative operation time corresponding to the current operation event number is subtracted from the cumulative operation time corresponding to the previous operation event number to obtain the event execution time; wherein, the previous operation event number is determined according to the order of cumulative operation time in the simulation platform interaction log, and the previous operation event number is the most recent operation event number in the simulation scenario dataset that belongs to the same trainee number and the same position number as the current operation event number, and whose cumulative operation time is less than the cumulative operation time corresponding to the current operation event number; the cumulative operation time corresponding to the previous operation event number is used to represent the occurrence time of the most recent operation before the current operation event, and the event execution time is used to represent the interval between the current operation event and the previous operation event. If there is no previous operation event number, the corresponding cumulative operation time is used as the event execution time. The absence of a previous operation event number indicates that the current event to be executed is the initial handling event, and its execution time is determined from the start time of this round of simulation. The event execution time is compared with the standard completion time limit corresponding to the current event to be executed. The standard completion time limit is the maximum allowed completion time pre-configured in the simulation rule base. When the event execution time exceeds the standard completion time limit, the current event to be executed is marked as a timed-out execution event. For timed-out execution events, the event execution time is subtracted from the standard completion time limit to obtain the event timeout duration. The event timeout duration indicates the specific length of time the current event to be executed exceeds the standard time limit and serves as the data basis for subsequent simulation evaluation and anomaly statistics.

[0042] In this implementation plan, the interval between adjacent operations is determined by the cumulative operation time sequence in the interaction log, and the timeout execution event is identified accordingly. This avoids evaluating solely based on whether the operation is completed and allows for further assessment of the trainee's response timeliness during the handling process. At the same time, by comparing the event execution time with the event standard completion time limit, the handling delay link can be quantitatively located, providing a clear timeliness basis for subsequent simulation evaluation, anomaly statistics, and training improvement.

[0043] Specifically, the process of generating event detection results based on abnormal event markers and statistically analyzing the number of abnormal and completed events is as follows: When an event to be executed does not have any abnormal event markers, it is determined to be a completed event. A completed event indicates that the event to be executed has been triggered by the current employee and has not been marked as an unexecuted event, an unsuccessful execution event, a timed-out execution event, or a skipped-level execution event. When the same event to be executed simultaneously meets two or more abnormal event markers, all corresponding abnormal event markers are retained. Multiple abnormal event markers indicate that the same event to be executed may simultaneously have both process non-compliance and timeliness non-compliance, i.e., simultaneously have timed-out execution event and skipped-level execution event markers, avoiding omission of one type of abnormality when abnormality types overlap. The current list of events to be executed is read and associated with the corresponding abnormal event markers and event timeout durations to form event detection results. Event detection results are recorded in units of events to be executed, and each event detection result includes at least the standard handling event number, responsible position number, preceding event number, event importance level, abnormal event marker, event timeout duration, and completion determination result. The event timeout duration is represented by a standardized numerical field, with the time unit consistent with the event execution time, preferably in seconds. When an event to be executed is marked as a timed-out event, the event timeout duration is the event execution time minus the event's standard completion time limit. When an event to be executed is not marked as a timed-out event, the event timeout duration is recorded as 0. Abnormal event markings are represented by a standardized enumeration field, including at least non-executed events, unsuccessful execution events, timed-out execution events, and out-of-level execution events. By using a standardized field format for event timeout durations and abnormal event markings, consistent data caliber is ensured when reading event detection results in subsequent abnormality count statistics, deduced completion scores, and process standardization scores. The event detection results are categorized according to abnormal event labels. Events marked as unexecuted, unsuccessful execution, timed-out, or bypassed execution are written into a critical event anomaly list. This critical event anomaly list records events that did not meet standard handling requirements in this training exercise and their reasons for the anomalies. Events judged as completed are written into a completed event list, which records events that met standard handling requirements in this training exercise. The number of unexecuted events, unsuccessful execution events, timed-out, bypassed execution events, and completed events are counted separately. These numbers are then summed to obtain the number of abnormal events. The number of abnormal events represents the total number of events marked with anomalies in this training exercise, while the number of completed events represents the total number of events that met standard handling requirements in this training exercise.The number of abnormal events represents the total number of abnormal event markers generated in this round of training simulation. When the same actionable event has two or more abnormal event markers, each marker is counted separately. Therefore, the number of abnormal events is allowed to be greater than the number of duplicate actionable events with abnormal markers. When the same actionable event has two or more abnormal event markers, deduplication is not performed; instead, each marker is counted separately. This is because different abnormal event markers correspond to different dimensions of abnormal handling. Deduplication would result in the loss of some abnormality type information, failing to accurately reflect the occurrence of various abnormal issues. Therefore, abnormal event markers are used as the statistical object, and different markers are counted separately to retain the distribution characteristics of abnormality types, providing a basis for subsequent simulation evaluation, identification of weak training events, and identification of key improvement positions. The number of completed events represents the number of actionable events that do not have any abnormal event markers and meet the standard handling requirements. Event detection results, a list of key event abnormalities, and a list of completed events are written into the training simulation management database for subsequent simulation evaluation, historical abnormality statistics, identification of weak training events, and identification of key improvement positions. The critical event exception list and the completed event list include statistical result fields for the number of non-executed events, the number of unsuccessful execution events, the number of timed-out execution events, the number of over-level execution events, and the number of completed events.

[0044] In this embodiment, Table 1 is a critical event anomaly detection data table. The table records in detail the standard handling event number, event name, preceding event number, standard completion time limit, operation event number, cumulative operation time, operation status, event execution time, and event flag for a trainee in a fire response post during a simulation. Among them, the standard handling event number E01 corresponds to the event name "Confirm Fire", the preceding event number is "No preceding event", the standard completion time limit for the event is 60, the operation event number is E01, the cumulative operation time is 45, the operation status is "Successfully Executed", the event execution time is 45, and the event is marked as "Completed Event"; the standard handling event number E02 corresponds to the event name "Activate Fire Broadcast", the preceding event number is E01, the standard completion time limit for the event is 90, the operation event number is E02, the cumulative operation time is 160, the operation status is "Successfully Executed", the event execution time is 115, and the event is marked as "Timeout Execution Event"; the standard handling event number E03 corresponds to the event name "Organize Evacuation of Key Areas", the preceding event number is E02, the standard completion time limit for the event is 120, the operation event number is E01, the cumulative operation time is 45, the operation status is "Successfully Executed", the event execution time is 115, and the event is marked as "Timeout Execution Event". The event number is Null, the cumulative operation time is Null, the operation status is Null, the event execution time is Null, and the event is marked as an unexecuted event; the standard handling event number E04 corresponds to the event name "Closing Fireproof Roller Shutter", the preceding event number is E02, the standard completion time limit for the event is 80, the operation event number is E04, the cumulative operation time is 210, the operation status is execution failure, the event execution time is 50, and the event is marked as an execution unsuccessful event; the standard handling event number E05 corresponds to the event name "On-site Verification Confirmation", the preceding event number is E03, the standard completion time limit for the event is 100, the operation event number is E05, the cumulative operation time is 260, the operation status is execution successful, the event execution time is 50, and the event is marked as an overstepping execution event.

[0045] Table 1. Key Event Anomaly Detection Data Table

[0046]

[0047] like Figure 3 The diagram illustrates the standard handling procedure and key event anomaly detection results. It demonstrates the process of detecting and marking the execution status of each standard handling event according to the standard handling procedure sequence in a training simulation scenario of an initial fire in a commercial complex. The horizontal arrows in the diagram represent the standard handling procedure sequence, with each node corresponding to standard handling events E01 to E05. The standard completion time limit for the corresponding event is marked below each node, and the actual cumulative operation time and detection results are marked above each node. By reading the cumulative operation time, operation status, and completion status of preceding events from the operation log, the completion status, timeliness, and procedure compliance of each standard handling event are detected, and the corresponding event marking results are output. Based on Table 1 and... Figure 3It can be seen that only incident E01 was completed in a timely manner according to the standard procedure requirements during this training simulation. Several typical abnormal situations occurred in the subsequent handling process. Although the operation was completed in E02, the completion time significantly exceeded the prescribed time limit, indicating that the trainees' response speed was insufficient after the incident escalated. E03 was not executed at all, indicating that a key evacuation step was omitted. Although E04 was triggered, the execution failed, indicating that the trainees were aware of the handling step but did not complete it correctly. Finally, E05 was executed directly without the completion of the preceding incident E03, forming a skipped-level execution phenomenon, indicating that the trainees did not strictly follow the standard handling procedure sequence to carry out emergency response.

[0048] In this implementation plan, by classifying and summarizing the abnormal event tags, it is possible to clearly distinguish the event statuses such as not executed, execution unsuccessful, execution timed out, execution exceeding level, and completed, so as to avoid confusion or omission of abnormal causes. At the same time, the event detection results, the list of critical event anomalies, and the list of completed events are uniformly written into the database, providing a structured and traceable data basis for subsequent simulation evaluation, historical statistics, and closed-loop improvement.

[0049] Specifically, combining the statistical results of the number of abnormal events and completed events with the simulation scenario dataset, the process of comprehensively evaluating the effectiveness of this round of simulation and forming a simulation evaluation record is as follows: The process involves reading the number of unexecuted events, unsuccessfully executed events, timed-out events, and out-of-level execution events from the critical event anomaly list, along with the event importance level corresponding to each anomaly. The number of unexecuted events, unsuccessfully executed events, timed-out events, and out-of-level execution events characterizes the distribution of different types of abnormal event handling in this training simulation, while the event importance level distinguishes the degree of impact of different abnormal events on the simulation evaluation results. Simultaneously, the process involves reading the number of completed events and the event importance level corresponding to each completed event from the completed event list. The number of completed events characterizes the number of events that meet the standard handling requirements in this training simulation, while the event importance level characterizes the importance of completed events within the standard handling process. The importance levels of all currently actionable events are summed to obtain the total importance level of actionable events. This total importance level represents the total importance of all standard handling events that trainees in the current position should complete in this training simulation. The importance level is represented numerically, preferably from level 1 to level 5. The importance levels of all completed events are summed to obtain the total importance level of completed events. This total importance level represents the total importance of handling events that trainees in the current position have completed according to standards in this training simulation. The importance levels of all overdue and bypassed events are summed to obtain the total importance level of process abnormal events. This total importance level represents the total importance of handling events with timeliness and process sequence abnormalities in this training simulation. The total importance level of completed events is divided by the total importance level of actionable events, and then multiplied by one hundred to obtain the simulation completion score. This simulation completion score characterizes the overall completion level of the corresponding actionable events by trainees in the current position. The process anomaly deduction value is obtained by dividing the total importance level of process anomalies by the total importance level of events requiring execution, and then multiplying by 100. Subtracting this deduction value from 100 yields the process standardization score. This score characterizes the degree of standardization in terms of timeliness and process sequence when trainees in the current position execute standard procedures. When the process standardization score is less than zero, it is corrected to zero to avoid negative scores due to the accumulation of multiple process anomalies. Multiplying by 100 converts the ratio to a percentage score, ensuring the final score ranges from 0 to 100, facilitating horizontal comparison and evaluation of results from different trainees, positions, and incident types.The current accident type, trainee number, job number, standard handling event number corresponding to the event to be executed, list of critical event anomalies, list of completed events, simulation completion score and process standardization score are linked and stored to form a simulation evaluation record, which is then written into the training simulation management database. This allows for subsequent tracking of the completion status, cause of anomalies and evaluation results of this round of training simulations according to trainees, jobs, accident types and standard handling events.

[0050] In this implementation plan, completed events and abnormal events are comprehensively evaluated by the importance level of the events. This can reflect the degree of event completion and process standardization of trainees in the current position, avoiding rough scoring based solely on the number of events. At the same time, the accident type, trainee, position, event list and scoring results are stored together, making the simulation evaluation results traceable, interpretable and reproducible, providing a data foundation for subsequent historical analysis and closed-loop improvement.

[0051] Specifically, the process of identifying weak training events and key improvement positions by combining historical simulation and evaluation records, and generating closed-loop improvement tasks, is as follows: Based on historical simulation and evaluation records, a list of historical critical event anomalies that has been written into the training simulation management database within a preset historical statistical period is retrieved. The preset historical statistical period is pre-set according to the training assessment cycle, preferably using the statistical range corresponding to the most recent 30 to 200 simulation records. The list of historical critical event anomalies is classified and statistically analyzed according to the standard handling event number, and the number of historical anomaly markings corresponding to each standard handling event is calculated. The number of historical anomaly markings is the cumulative number of times the same standard handling event number has been marked as an unexecuted event, an unsuccessful execution event, an overdue execution event, and an over-level execution event within the preset historical statistical period. When the number of anomaly markings corresponding to the same standard handling event is greater than the event anomaly threshold, the corresponding standard handling event number is marked as a weak training event. The event anomaly threshold is pre-determined based on the number of simulations within the preset historical statistical period, training assessment requirements, and the number of anomalies allowed by management personnel. Weak training events are used to represent standard handling events that repeatedly exhibit anomalies in historical simulations. The historical critical event anomaly list is categorized and statistically analyzed according to the responsibility position number, and the number of anomalies corresponding to each responsibility position is calculated separately. The number of anomalies corresponding to a responsibility position is the cumulative number of unexecuted events, unsuccessful execution events, timed-out execution events, and out-of-level execution events associated with the same responsibility position number within a preset historical statistical period. When the number of anomalies corresponding to the same responsibility position exceeds an anomaly number threshold, the corresponding responsibility position is marked as a key improvement position. The anomaly number threshold is pre-determined based on the number of times the corresponding responsibility position participated in simulations within the preset historical statistical period, the position's performance evaluation standards, and the anomaly tolerance number set by management personnel. Key improvement positions are used to indicate responsibility positions where anomalies occur frequently in historical simulations. A closed-loop improvement task list is generated and output for key improvement positions and weak training events. The closed-loop improvement task list includes key improvement positions, weak training events, corresponding anomaly event types, and task generation time, used to guide subsequent targeted training arrangements and simulation content optimization. Optionally, the closed-loop improvement task list may also include the responsible position, suggested improvement measures, task status, and completion confirmation information. The responsible position is used to determine the entity responsible for executing the corresponding improvement task. Suggested improvement measures are automatically generated based on the abnormal event type corresponding to the weak training event or configured by management personnel, including adding specialized training, supplementing process drills, strengthening timeliness training, or enhancing job collaboration training. After completing the improvement task, the corresponding responsible position submits a completion confirmation, recording the confirmation time and completion status.During subsequent retraining simulations, the system automatically reads the latest simulation evaluation records associated with the corresponding responsible positions and weak training events, comparing and verifying the number of abnormal events, simulation completion scores, and process standardization scores before and after improvement. When the number of abnormal events decreases and the simulation completion score or process standardization score meets the preset improvement requirements, the corresponding closed-loop improvement task is marked as completed; otherwise, the corresponding closed-loop improvement task is retained and continues to participate in subsequent training tracking. The verification results are written to the training simulation management database to update historical simulation evaluation records and serve as data for identifying key improvement positions and weak training events in the future.

[0052] In this implementation plan, by reading the list of historical key events anomalies within a preset historical statistical period, the abnormal situations of standard handling events and responsible positions are classified and statistically analyzed. This can identify weak training events that have repeatedly made mistakes over a long period of time and key improvement positions where anomalies occur in a concentrated manner, avoiding one-sided judgments based solely on the results of a single simulation. At the same time, a closed-loop improvement task list is generated based on the identification results, providing a basis for continuous improvement for subsequent targeted training arrangements, job capability enhancement, and simulation content optimization.

[0053] Reference Figure 2 As shown, the second aspect of the present invention provides a training simulation management system based on scenario data fusion, applied to the aforementioned training simulation management method based on scenario data fusion, comprising: a simulation data acquisition and processing module, used to acquire simulation scenario data in real time during the training simulation process, and preprocess the simulation scenario data to obtain a simulation scenario dataset; an execution event identification module, used to identify the execution events corresponding to the current training simulation based on the simulation scenario dataset, and construct an execution event list; a key event anomaly detection module, used to combine the execution event list and the simulation scenario dataset to analyze the completion status, process compliance, and timeliness of the execution events, mark anomalies, form event detection results based on the anomaly event markings, and perform statistical analysis of the number of anomalies and completed events; and a simulation evaluation closed-loop management module, used to combine the statistical results of the number of anomalies and completed events with the simulation scenario dataset to comprehensively evaluate the effect of this round of simulation, and form a simulation evaluation record; and to combine historical simulation evaluation records to identify weak training events and key improvement positions, and generate closed-loop improvement tasks.

[0054] like Figure 4The diagram illustrates the event-driven simulation evaluation process. It showcases the relationships between core components in a training simulation management system based on scenario data fusion. First, standard handling events, preceding event relationships, standard event completion deadlines, and event importance levels are obtained from the standard handling process. Combined with the current accident type and job information, the events to be executed in the current training simulation, which should be completed by the corresponding job position, are identified. Simultaneously, the operation event numbers, operation times, and operation statuses of trainees during the simulation are read from the actual operation logs to reflect their actual operational behavior. Then, using the events to be executed as the verification benchmark, they are analyzed against the actual operation logs, including non-execution detection, unsuccessful execution detection, timeout detection, and overstepping detection: non-execution detection determines whether there are critical events that should be executed but have not been triggered; unsuccessful execution detection identifies events that have been triggered but whose operation status has not reached successful execution; timeout detection determines whether executed events have exceeded the standard event completion deadline; and overstepping detection determines whether subsequent events have been executed prematurely when preceding events have not been completed. All anomaly detection results are ultimately compiled into the simulation and evaluation stage to form a list of critical event anomalies, a list of completed events, and an evaluation of the completion and standardization of subsequent simulations. Through a chain-like process of "standard handling procedures, events to be executed, actual operation logs, anomaly detection, and simulation evaluation," comprehensive identification of omissions, delays, and abnormalities in the sequence of critical handling events is achieved, thereby improving the accuracy, interpretability, and traceability of training simulation and evaluation results.

[0055] This implementation plan, through modular collaboration of simulation data collection and processing, execution event identification, key event anomaly detection, and simulation evaluation closed-loop management, enables unified processing of simulation scenario data, accurate identification of execution events, classification and detection of anomalies, and continuous tracking of historical weaknesses. This improves the completeness, interpretability, and traceability of training simulation evaluation, and provides data support for subsequent targeted training and job improvement.

[0056] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0057] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. As those skilled in the art will understand, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A training simulation management method based on scenario data fusion, characterized in that, Includes the following steps: S1 collects simulation scenario data in real time during the training simulation process and preprocesses the simulation scenario data to obtain the simulation scenario dataset; S2, based on the simulation scenario dataset, identify the events that should be executed corresponding to the current training simulation and construct a list of events that should be executed; S3 combines the list of events to be executed with the simulation scenario dataset to analyze the completion status, process compliance, and timeliness of the events to be executed, marks abnormal events, generates event detection results based on the abnormal event markings, and performs statistical analysis on the number of abnormal events and completed events; S4, combining the statistical results of the number of abnormal events and completed events with the simulation scenario dataset, comprehensively evaluates the effect of this round of simulation and forms a simulation evaluation record; combining the historical simulation evaluation records, identifies weak training events and key improvement positions, and generates closed-loop improvement tasks.

2. The training simulation management method based on scenario data fusion according to claim 1, characterized in that, The specific process of real-time collection of simulation scenario data during training simulation is as follows: Real-time acquisition of simulation scenario data, specifically including: Read the current incident type from the currently loaded simulation script, and read the simulation start time from the simulation platform system clock; Read the student ID, job ID, operation event ID, operation time, and operation status from the simulation platform's interaction log. The operation status includes successful execution and failed execution. Read standard handling process data from the simulation rule base, including standard handling event number, applicable accident type, responsible position number, preceding event number, event standard completion time limit, and event importance level.

3. The training simulation management method based on scenario data fusion according to claim 2, characterized in that, The specific process of preprocessing the simulation scenario data to obtain the simulation scenario dataset is as follows: The simulation scenario data is converted into a unified encoding format and a unified time unit; the start time of this simulation round is used as the unified time zero point, and the start time of this simulation round is subtracted from the start time of each operation time to obtain the corresponding cumulative operation time; for duplicate data records with the same student number, job number, and operation event number, only the record with the largest cumulative operation time is retained; when the operation status is empty, the operation status is marked as not returned; when the preceding event number is empty, the preceding event number is marked as no preceding event, otherwise it is marked as having a preceding event; Based on the preprocessed simulation scenario data, a simulation scenario dataset is formed and written into the training simulation management database.

4. The training simulation management method based on scenario data fusion according to claim 3, characterized in that, The specific process of identifying the events to be executed corresponding to the current training simulation based on the simulation scenario dataset and constructing a list of events to be executed is as follows: The current accident type is compared with the applicable accident types in the standard handling procedure data, and the current job number is compared with the responsible job number in the standard handling procedure data; when the applicable accident type is the same as the current accident type and the responsible job number is the same as the current job number, the corresponding standard handling event number is marked as the event to be executed. Read the preceding event number corresponding to the event to be executed, and form a standard handling process sequence based on the relationship between the preceding event numbers, which serves as the basis for identifying subsequent over-level execution events; Read the responsible position number, preceding event number, standard completion deadline and importance level of each event to be executed, associate and store them to form a list of events to be executed, and write it into the training simulation management database.

5. The training simulation management method based on scenario data fusion according to claim 4, characterized in that, The specific process of combining the list of events to be executed with the simulation scenario dataset to analyze the completion status, process compliance, and timeliness of the events to be executed, and to mark abnormal events, is as follows: For each event to be executed, an abnormal event is marked: the standard handling event number corresponding to the event to be executed is compared with the operation event number in the simulation scenario dataset; If there is no operation event number in the simulation scenario dataset that is the same as the standard handling event number, the event that should be executed will be marked as an unexecuted event. When there is an operation event number in the simulation scenario dataset that is the same as the standard handling event number, read the operation status corresponding to the operation event number. When the operation status is not "execution successful", mark the event that should be executed as "execution unsuccessful"; Read the preceding event number corresponding to each event to be executed. When the preceding event number is marked as having a preceding event, determine whether there is an operation event number with the same preceding event number in the simulation scenario dataset, and whether the corresponding operation status is successful. If the conditions are not met and the operation status of the operation event number corresponding to the current event to be executed is successful, then mark the current event to be executed as a skip-level execution event. At the same time, for each event that should be executed, timeout execution events are identified.

6. The training simulation management method based on scenario data fusion according to claim 5, characterized in that, The specific process for identifying timeout execution events for each event that should be executed is as follows: For each event to be executed, if there is an operation event number in the simulation scenario dataset that is the same as the standard handling event number, and the operation status is successful, read the cumulative operation time corresponding to the successful operation event number, and subtract the cumulative operation time corresponding to the previous operation event number from the cumulative operation time corresponding to the current operation event number to obtain the event execution time. The previous operation event number is determined according to the order of cumulative operation times in the simulation platform interaction log; if there is no previous operation event number, the corresponding cumulative operation time is used as the event execution time. The execution time of the event is compared with the standard completion time limit of the event corresponding to the current event to be executed. When the execution time of the event is greater than the standard completion time limit, the current event to be executed is marked as a timed-out execution event. For timed-out execution events, the execution time of the event is subtracted from the standard completion time limit to obtain the event timeout duration.

7. The training simulation management method based on scenario data fusion according to claim 6, characterized in that, The specific process of generating event detection results based on abnormal event tags and performing statistical analysis on the number of abnormal events and completed events is as follows: If no exception event flag is present for an event to be executed, it is considered a completed event. If the same event to be executed satisfies two or more exception event flags, all corresponding exception event flags are retained. Read the current list of events to be executed, and associate it with the corresponding exception event flags and event timeout durations to form event detection results; The event detection results are classified according to the abnormal event labels. Events that should be executed and are marked as non-executed events, execution unsuccessful events, timed-out execution events, and out-of-level execution events are written into the critical event exception list. Events that should be executed and are judged as completed events are written into the completed event list. The number of unexecuted events, unsuccessful execution events, timed-out execution events, out-of-level execution events, and completed events are counted separately. The number of unexecuted events, unsuccessful execution events, timed-out execution events, and out-of-level execution events are then added together to obtain the number of abnormal events. The event detection results, the list of critical event anomalies, and the list of completed events are written into the training simulation management database.

8. The training simulation management method based on scenario data fusion according to claim 7, characterized in that, The specific process of combining the statistical results of the number of abnormal events and completed events with the simulation scenario dataset to comprehensively evaluate the effectiveness of this round of simulation and form a simulation evaluation record is as follows: Read the number of non-executed events, the number of unsuccessful execution events, the number of timed-out execution events, and the number of over-level execution events from the critical event exception list, as well as the event importance level corresponding to the exception events; Simultaneously read the number of completed events and the importance level of each completed event from the completed event list; Sum the importance levels of all events that should be executed to obtain the total importance level of the events that should be executed; Sum the importance levels of all completed events to obtain the total importance level value of completed events; Sum the importance levels of all timed-out and overstepped execution events to obtain the total importance level of process exception events; divide the total importance level of completed events by the total importance level of events to be executed, and multiply by one hundred to obtain the simulation completion score; Divide the total importance level of process exception events by the total importance level of events to be executed, and multiply by one hundred to obtain the process exception deduction value. Subtract the process exception deduction value from one hundred to obtain the process standardization score. When the process standardization score is less than zero, the process standardization score is corrected to zero. The current accident type, trainee number, job number, standard handling event number corresponding to the event to be executed, list of critical event anomalies, list of completed events, simulation completion score and process standardization score are linked and stored to form a simulation evaluation record, which is then written into the training simulation management database.

9. The training simulation management method based on scenario data fusion according to claim 8, characterized in that, The specific process of combining historical simulation and evaluation records to identify weak training events and key improvement positions, and generating closed-loop improvement tasks is as follows: Based on historical simulation and evaluation records, the list of historical key events is classified and statistically analyzed according to the standard handling event number, and the number of historical anomaly markers corresponding to each standard handling event is calculated. When the number of anomaly markers corresponding to the same standard handling event exceeds the event anomaly threshold, the corresponding standard handling event number will be marked as a weak training event. The historical critical event anomaly list is classified and statistically analyzed according to the responsibility position number, and the number of anomalies corresponding to each responsibility position is calculated. When the number of anomalies corresponding to the same responsibility position is greater than the anomaly number threshold, the corresponding responsibility position is marked as a key improvement position. For key improvement positions and weak training events, generate and output a closed-loop improvement task list.

10. A training and simulation management system based on scenario data fusion, characterized in that, include: The simulation data acquisition and processing module is used to collect simulation scenario data in real time during the training simulation process, and to preprocess the simulation scenario data to obtain the simulation scenario dataset. The Executable Event Identification Module is used to identify the Executable Events corresponding to the current training simulation based on the simulation scenario dataset, and to build a list of Executable Events; The critical event anomaly detection module is used to combine the list of events to be executed with the simulation scenario dataset to analyze the completion status, process compliance, and timeliness of the events to be executed, mark anomaly events, generate event detection results based on the anomaly event markings, and perform statistical analysis on the number of anomaly events and completed events. The simulation evaluation closed-loop management module is used to comprehensively evaluate the effect of this round of simulation by combining the statistical results of the number of abnormal events and completed events with the simulation scenario dataset, and form a simulation evaluation record; combined with historical simulation evaluation records, it identifies weak training events and key improvement positions, and generates closed-loop improvement tasks.

Images