A virtual reality-based intelligent emergency drill method and system for chemical enterprises

Abstract

The present invention discloses a virtual reality-based intelligent emergency drill method and system for chemical enterprises, which relates to the field of virtual reality emergency management technology. The method comprises the following steps: collecting individual behavior paths, historical accident data, and chemical production process parameters; constructing a multivariate fault evolution knowledge graph based on the historical accident data and chemical production process parameters to obtain a multidimensional coupling matrix; inputting the multidimensional coupling matrix into a virtual reality scene generation unit to obtain initial drill data; constructing a behavioral event sequence based on the initial drill data to obtain an operation trajectory data set; and converting the operation trajectory data set into a directed behavior path. The present invention matches a set of error pattern templates with individual behavior paths to obtain a set of scenario replay parameters, thereby achieving automatic identification and parameterized extraction of behavioral deviations and error types, and dynamically adjusting them in combination with individual paths.

Description

Technical Field

[0001] The present invention relates to the technical field of virtual reality emergency management, and in particular to a virtual reality-based intelligent emergency drill method and system for chemical enterprises. Background Art

[0002] In the safety management practice of chemical companies, emergency drill mechanisms based on historical case analysis and operating procedures are an important means to improve personnel's emergency response capabilities. Conventional emergency drill methods mainly include three forms: desktop simulation, on-site simulation, and virtual reality-assisted simulation. Their core purpose is to restore accident scenarios and simulate operating procedures to test and improve the judgment ability and handling efficiency of operators when facing emergencies. Virtual reality technology has been gradually introduced into emergency drill systems to build three-dimensional visualization environments and interactive processes, and combined with preset accident models, evacuation routes and alarm mechanisms to achieve the reproduction of typical accident scenarios. Existing methods mostly rely on fixed process modeling and static script driving, lacking deep integration of complex chemical processes, individual behavior differences and accident evolution mechanisms, and difficult to dynamically reflect the coupling variables and personnel behavior deviations in actual accidents.

[0003] However, from the perspective of intelligent drills, the existing virtual drill methods based on fixed processes and predefined scenarios have two key limitations: first, it is difficult to combine individual historical operational behaviors with accident evolution paths to construct dynamic scenario content, making it impossible to achieve behavior-driven drill customization; second, in the process of scenario generation and interaction logic construction, there is a lack of support mechanisms based on knowledge graphs and multivariate path mapping, resulting in the complexity, personalization and realism of the drill content, which limits its adaptability under high-risk working conditions. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a virtual reality-based intelligent emergency drill method for chemical enterprises to solve the problems that the drill process is difficult to dynamically generate based on individual behavior and lacks a multivariable evolution support mechanism.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a virtual reality-based intelligent emergency drill method for chemical enterprises, which includes collecting individual behavior paths, historical accident data, and chemical production process parameters, constructing a multivariate fault evolution knowledge graph based on the historical accident data and chemical production process parameters, and obtaining a multidimensional coupling matrix;

[0008] Input the multi-dimensional coupling matrix into the virtual reality scene generation unit to obtain the initial data of the drill;

[0009] Based on the initial data of the drill, a behavioral event sequence is constructed to obtain a set of operation trajectory data;

[0010] The operation trajectory data set is converted into a directed behavior path, and the directed behavior path is identified through a graph mining algorithm to obtain a set of error pattern templates of the directed behavior path;

[0011] The error pattern template set is matched with the individual behavior path to obtain the scene replay parameter set. According to the scene replay parameter set, a personalized virtual reality emergency process is generated.

[0012] As a preferred solution of the virtual reality-based intelligent emergency drill method for chemical enterprises described in the present invention, the following steps are included: collecting individual behavior paths, historical accident data, and chemical production process parameters, constructing a multivariate fault evolution knowledge graph based on the historical accident data and chemical production process parameters, and obtaining a multidimensional coupling matrix.

[0013] Collect the operational behavior sequences and response decision data generated by trainees in different emergency drill tasks to form individual behavior paths;

[0014] Organize the chemical enterprise accident case database and extract chemical production process parameters from the equipment layout of chemical enterprises;

[0015] Utilize graph structure modeling methods to construct a multivariate fault evolution knowledge graph from historical accident data and chemical production process parameters;

[0016] Based on the multivariable fault evolution knowledge graph, it is mapped into a multidimensional coupling matrix.

[0017] As a preferred solution of the virtual reality-based intelligent emergency drill method for chemical enterprises of the present invention, wherein: a multi-dimensional coupling matrix is ​​input into a virtual reality scene generation unit to obtain initial drill data, comprising the following steps:

[0018] The accident cause combinations and variable evolution path information are extracted from the multidimensional coupling matrix. Based on the accident cause combinations, the spatial coordinates of the fault source are located in the three-dimensional equipment model. Based on the variable evolution path information, the diffusion range of hazardous substances is obtained through physical modeling.

[0019] Based on the diffusion range of dangerous substances and individual behavior paths, the evacuation path point set is obtained;

[0020] According to the variable evolution path information, set the animation response logic for the 3D model;

[0021] The diffusion range of hazardous materials, evacuation route point set, animation response logic and fault source spatial coordinates are input into the virtual reality scene to generate initial data for the drill.

[0022] As a preferred solution of the virtual reality-based intelligent emergency drill method for chemical enterprises of the present invention, wherein: based on the initial drill data, a behavior event sequence is constructed to obtain an operation trajectory data set, including the following steps:

[0023] Read the initial data of the drill and obtain the set of interactive objects;

[0024] Perform task mapping on the interactive object set, generate a standard operation event list, record the trainees' operations in the virtual reality environment, and obtain the original behavior action flow;

[0025] Perform event semantic annotation on the standard operation event list and the original behavior action flow to form a preliminary behavior event sequence;

[0026] The preliminary behavior event sequence is subjected to trajectory extraction to obtain a set of behavior trajectory fragments, which are then aggregated by individuals to obtain an operation trajectory data set.

[0027] As a preferred solution of the virtual reality-based intelligent emergency drill method for chemical enterprises of the present invention, the operation trajectory data set is converted into a directed behavior path, the directed behavior path is identified by a graph mining algorithm, and an error pattern template set of the directed behavior path is obtained, including the following steps:

[0028] Extract action nodes and time sequence relationships from the operation trajectory data set and construct the initial directed graph of operation events;

[0029] Sort the initial directed graph of operation events by the operation time sequence to obtain a time series directed behavior path graph;

[0030] Based on the frequent subgraph mining algorithm, the pattern of the time series directed behavior path graph is extracted to identify the frequent behavior subpath set;

[0031] Filter the frequent behavior sub-path set to obtain the sub-path set that deviates from the standard path;

[0032] The set of subpaths that deviate from the standard path are semantically classified and annotated with error labels to obtain a set of error pattern templates for directed behavior paths.

[0033] As a preferred solution of the virtual reality-based intelligent emergency drill method for chemical enterprises of the present invention, the error pattern template set is matched with the individual behavior path to obtain the scene replay parameter set, including the following steps:

[0034] Format the individual behavior path according to the node number and time sequence to obtain the individual behavior path data sequence, encode the node pattern and edge relationship of each error path template in the error pattern template set, and establish the error path template encoding index table;

[0035] Use the subgraph isomorphism matching algorithm to match the individual behavior path data sequence with the error path template encoding index table to obtain a matching result list;

[0036] Based on the matching result list, trigger conditions and timing information, a set of behavioral deviation fragments is generated, and the deviation type identified in the set of behavioral deviation fragments is used to obtain a preliminary scene replay parameter set;

[0037] The preliminary scene replay parameter set and the individual behavior path are dynamically adjusted to obtain the scene replay parameter set.

[0038] As a preferred solution of the virtual reality-based intelligent emergency drill method for chemical enterprises of the present invention, a personalized virtual reality emergency process is generated according to a set of scene replay parameters, including the following steps:

[0039] Parse the behavior deviation information, trigger conditions and scene position parameters in the scene replay parameter set to generate an event-driven script configuration table;

[0040] Based on the event-driven script configuration table, a 3D scene object set is generated using the 3D scene component library;

[0041] Based on the 3D scene object set, a drill process timing model is established;

[0042] By rehearsing the process sequence model, we can obtain the configuration table of the interaction logic mechanism;

[0043] Organize the 3D scene object collection and interactive logic mechanism configuration table to generate emergency process data;

[0044] Deploy emergency process data to the drill platform to generate personalized virtual reality emergency processes.

[0045] In a second aspect, the present invention provides a virtual reality-based intelligent emergency drill system for chemical enterprises, including a multidimensional coupling matrix module that collects individual behavior paths, historical accident data, and chemical production process parameters, and constructs a multivariate fault evolution knowledge graph based on the historical accident data and chemical production process parameters to obtain a multidimensional coupling matrix;

[0046] The drill initial data module inputs the multi-dimensional coupling matrix into the virtual reality scene generation unit to obtain the drill initial data;

[0047] The operation trajectory data collection module constructs a behavioral event sequence based on the initial drill data to obtain the operation trajectory data collection;

[0048] The error pattern template module converts the operation trajectory data set into a directed behavior path, identifies the directed behavior path through a graph mining algorithm, and obtains the error pattern template set of the directed behavior path;

[0049] The virtual reality emergency process module matches the error pattern template set with the individual behavior path to obtain the scene replay parameter set, and generates a personalized virtual reality emergency process based on the scene replay parameter set.

[0050] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the virtual reality-based intelligent emergency drill method for chemical enterprises as described in the first aspect of the present invention is implemented.

[0051] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the virtual reality-based intelligent emergency drill method for chemical enterprises as described in the first aspect of the present invention.

[0052] The beneficial effects of the present invention are as follows: by constructing a multivariable fault evolution knowledge graph and mapping it into a multidimensional coupling matrix, the causal variables, process flow parameters and accident evolution logic involved in chemical accidents are structured and integrated, so that the accident causal chain has visualization and computable characteristics in the graph, thereby providing a semantically rich and logically closed data foundation for the subsequent virtual rehearsal scene generation, thereby improving the pertinence and complexity expression ability of the rehearsal content, and by matching the error pattern template set with the individual behavior path, obtaining the scene replay parameter set, realizing the automatic identification and parameterized extraction of behavioral deviations and error types, and combining the individual path for dynamic adjustment, thereby driving the personalized replay configuration of the virtual scene, effectively enhancing the behavior-driven mechanism and emergency feedback adaptability of the rehearsal process. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0054] Figure 1 This is a flow chart of the intelligent emergency drill method for chemical enterprises based on virtual reality.

[0055] Figure 2 Schematic diagram of the intelligent emergency drill system for chemical enterprises based on virtual reality. DETAILED DESCRIPTION

[0056] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0057] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0058] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0059] Reference Figure 1~Figure 2 , is an embodiment of the present invention, which provides a chemical enterprise intelligent emergency drill method based on virtual reality, comprising the following steps:

[0060] S1. Collect individual behavior paths, historical accident data, and chemical production process parameters. Based on the historical accident data and chemical production process parameters, construct a multivariate fault evolution knowledge graph to obtain a multidimensional coupling matrix.

[0061] S1.1. Collect the operational behavior sequences and response decision data generated by trainees in different emergency drill tasks to form individual behavior paths.

[0062] Furthermore, in different types of emergency drill tasks, the trainees' operational responses to various emergencies in the virtual environment are recorded in real time, including button clicks, equipment operations, path selection, and information reading interaction behaviors. At the same time, the judgments and decision results made by the acquisition nodes are synchronously collected, and the corresponding timestamps, behavioral goals and execution sequences are marked. All operational behaviors and decision data are organized and numbered according to the task process, and a time-serialized behavioral chain is established. Individuals are classified to form individual behavioral paths.

[0063] S1.2. Organize the accident case database of chemical enterprises and extract the chemical production process parameters in the equipment layout of chemical enterprises.

[0064] Furthermore, typical accident cases that have occurred in chemical companies over the years are collected and classified, and key information such as the causes, processes, impact scope and disposal results of various accidents are sorted out. The focus is on identifying the types of equipment involved in the accident, their layout locations and their mutual connections. At the same time, chemical production process parameters related to the accident are extracted, including dynamic variables such as logistics transmission paths, temperature, pressure, material types and reaction timing between each device, and combined with the functional positioning of the equipment in the process flow, chemical production process parameters are formed.

[0065] S1.3. Use graph structure modeling methods to construct historical accident data and chemical production process parameters into a multivariate fault evolution knowledge graph.

[0066] Furthermore, the historical accident data and chemical production process parameters obtained are structured and coded, with accident causes, equipment units, process variables, evolution paths, and response behavior elements as nodes in the graph. The edge connection method is constructed according to the time sequence, logical causality and physical dependency relationship, forming a graph structure with multiple attributes and relationship types. By marking the triggering conditions, probability weights and variable change trends of each edge, a unified expression of the fault propagation path, variable coupling relationship and personnel response behavior is achieved, thereby constructing a multivariate fault evolution knowledge graph.

[0067] S1.4. Based on the multivariable fault evolution knowledge graph, it is mapped into a multidimensional coupling matrix.

[0068] Furthermore, based on the constructed multivariable fault evolution knowledge graph, we first extract the key elements such as accident causes, process equipment, variable changes and response behaviors corresponding to each node in the graph, and quantify the correlation between these elements. According to the connection strength, trigger probability and path weight between nodes, we construct a matrix mapping rule, map the various relationships in the graph to row and column indexes in the matrix, and fill in the corresponding coupling strength, evolution direction and time-related parameters with causes, equipment and variables as dimension axes. Finally, we form a multidimensional coupling matrix that can comprehensively reflect the fault diffusion process, multi-source data linkage and behavioral feedback correlation.

[0069] S2. Input the multi-dimensional coupling matrix into the virtual reality scene generation unit to obtain the initial data of the drill.

[0070] S2.1. Extract the accident cause combination and variable evolution path information from the multidimensional coupling matrix. Based on the accident cause combination, locate the spatial coordinates of the fault source in the three-dimensional equipment model. Based on the variable evolution path information, obtain the diffusion range of hazardous substances through physical modeling.

[0071] Furthermore, within the multidimensional coupling matrix, we first identify areas in the matrix that represent high coupling between triggers and equipment and process variables. We then screen out combinations with strong triggering associations between the trigger nodes and other elements. By combining the frequency of occurrence of triggers in historical accidents, their triggering paths, and the frequency of interaction with key equipment or operations, we extract common accident trigger combinations. During this extraction process, we retain the temporal sequence logic and spatial location distribution information between the triggers.

[0072] The accident cause combinations extracted from the multidimensional coupling matrix are correlated with the functional modules of the chemical equipment to determine the key components and specific device units that may cause the failure. Based on the spatial layout data of these components in the three-dimensional equipment model, the corresponding spatial coordinates of the fault source are located. Further combined with the parameters such as temperature, pressure, medium flow rate and reaction time involved in the variable evolution path information, a numerical simulation method based on mass conservation is constructed to calculate the spatiotemporal evolution process of the diffusion of hazardous substances from the fault source location to the outside, as well as the diffusion range of hazardous substances.

[0073] S2.2. Based on the diffusion range of hazardous substances and individual behavior paths, obtain the evacuation path point set.

[0074] Furthermore, based on the diffusion boundaries of hazardous substances in three-dimensional space, dangerous areas that are inaccessible to people at each moment are delineated. The initial position and historical travel direction recorded in the individual behavior path are used as the starting reference point. Combined with the spatial coordinates of safe passages, entrances and exits, and safe havens in the chemical scene, a passable area map is constructed. The time factor is introduced into this map, and the path pass weight is dynamically adjusted according to the dynamic evolution process of the dangerous area. A series of shortest or optimal pass node sequences that avoid dangerous areas and lead to safe target areas are extracted through the path search algorithm to obtain an evacuation path point set.

[0075] S2.3. Set the animation response logic for the three-dimensional model based on the variable evolution path information.

[0076] Furthermore, based on the temporal change trends of key variables (such as pressure, temperature, concentration, etc.) in the variable evolution path information, the physical change characteristics that the corresponding equipment or components may exhibit in different states are determined, and the physical change characteristics are bound to the corresponding equipment objects in the three-dimensional model. When the pressure variable reaches the threshold, dynamic effects such as pipeline expansion, rupture or valve opening are triggered; further, according to the temporal relationship of variable changes and the logical cause and effect between the evolution path nodes, the trigger conditions and execution order of each animation are set to ensure that different events have a reasonable response sequence and continuity in the three-dimensional space, thereby forming an animation response logic consistent with the accident evolution process.

[0077] S2.4. Input the hazardous material diffusion range, evacuation route point set, animation response logic, and fault source spatial coordinates into the virtual reality scene to generate initial drill data.

[0078] Furthermore, the boundaries and shapes of the affected areas in the virtual environment are first determined based on the spread range of hazardous substances, and are marked by example values ​​such as a circular area with a diffusion radius of 50 meters or an irregular polygonal area based on wind direction simulation; then, a complete evacuation route is constructed in the virtual reality scene in combination with the evacuation path point set. The point set consists of multiple pre-set geographic coordinate points, which are used to indicate the specific direction and path of personnel evacuation; then, the animation response logic is embedded in the virtual reality scene. The logic includes the switching of personnel behavior states, equipment action feedback, and the visualization of environmental changes, and drives the corresponding animation effects according to the preset time sequence and trigger conditions. Finally, the initial position of the accident is located according to the spatial coordinates of the fault source, and the specific spatial position of the fault source is marked in the virtual reality scene as the starting point of the hazardous substance spread range and the starting point reference of the evacuation path point set.

[0079] S3. Based on the initial drill data, a behavioral event sequence is constructed to obtain an operation trajectory data set.

[0080] S3.1. Read the initial data of the drill and obtain a set of interactive objects.

[0081] Furthermore, the spatial distribution information of the diffusion range of hazardous substances is combined with the spatial coordinates of the evacuation path point set, combined with the trigger conditions and dynamic performance of each device and environmental change in the animation response logic, and the spatial coordinates of the fault source are used to determine the starting position of the accident. These data are integrated as a whole to construct the basic environment and dynamic interactive elements of the virtual reality scene, and finally generate the initial drill data including the scene spatial layout, dynamic response mechanism, personnel evacuation path and the accident development process.

[0082] S3.2. Task mapping is performed on the set of interactive objects to generate a list of standard operation events. The trainees' operations in the virtual reality environment are recorded to obtain the original behavior action flow.

[0083] Furthermore, the functions of each object in the set of interactive objects are identified and associated with tasks, and its typical operation tasks in the emergency response process are mapped into specific action event items. Based on this, a standard operation event list containing object identification, task type and operation sequence is formed. In the virtual reality environment, every actual operation behavior of the trainees with these interactive objects is recorded in real time, including operation time, action type, interactive object and its state change, thereby forming a chronologically arranged original behavior action flow.

[0084] S3.3. Annotate the standard operation event list and the original behavior action flow with event semantics to form a preliminary behavior event sequence.

[0085] Furthermore, based on the operation task types, object characteristics and expected behaviors defined in the standard operation event list, each operation record in the original behavior action flow is compared and analyzed item by item, and combined with the operation timing and context action relationship, each operation event is given a corresponding semantic label, such as closing the valve and starting the pump to confirm the alarm. The labeling content includes the purpose of the behavior, the function of the object and the response category. Through this semantic labeling process, the original operation data is converted into a preliminary behavior event sequence with clear behavior meaning, thereby realizing the clarification and standardization of the behavior logic.

[0086] S3.4. Extract the trajectory of the preliminary behavior event sequence to obtain a set of behavior trajectory fragments, and aggregate the set of behavior trajectory fragments by individuals to obtain an operation trajectory data set.

[0087] Furthermore, in the marked preliminary behavioral event sequence, several behavioral segments with logical association and operational coherence are extracted according to the chronological order of the events and their continuity in spatial position, forming a set of behavioral trajectory fragments; each fragment represents the trainee's operational path and decision-making response process in a specific emergency task stage; then all behavioral trajectory fragments are classified and sorted according to their individual numbers, and aggregated into the complete operational trajectory of each trainee during the entire exercise process, thereby obtaining an operational trajectory data set.

[0088] S4. Convert the operation trajectory data set into a directed behavior path, identify the directed behavior path through a graph mining algorithm, and obtain a set of error pattern templates of the directed behavior path.

[0089] S4.1. Extract action nodes and time sequence relationships from the operation trajectory data set and construct an initial directed graph of operation events.

[0090] Furthermore, in the operation trajectory data set, for each trainee's trajectory, the key operation content in their behavior fragments is analyzed one by one, and each behavior fragment with clear operational significance is identified as an independent action node. The action nodes are sorted according to the actual occurrence time of each action node during the drill, and the execution sequence relationship between the nodes is marked. According to the time sequence and operation logic, they are connected through directed edges to form an initial directed graph of operation events.

[0091] S4.2. Sort the initial directed graph of operation events according to the operation time sequence to obtain a time sequence directed behavior path graph.

[0092] Furthermore, in the constructed initial directed graph of operation events, all nodes are globally sorted in chronological order based on the occurrence time recorded by each action node, keeping the original causal relationship between the nodes unchanged. Only for node combinations with concurrent or disordered connections, clear directed edges are re-established based on chronological order, ultimately forming a directed behavior path graph with clear time series and logical continuity.

[0093] S4.3. Based on the frequent subgraph mining algorithm, the pattern of the time series directed behavior path graph is extracted to identify the frequent behavior subpath set.

[0094] Furthermore, the behavioral paths of all individuals in the time series directed behavioral path graph are traversed, and the frequent subgraph mining algorithm is used to analyze the subgraph structure that appears repeatedly in multiple behavioral paths. By setting the minimum support threshold, the behavioral node combinations with high frequency of recurrence in multiple operation trajectories and their time sequence relationships are screened out, and the directed subpaths that meet the support requirements are extracted as frequent behavioral patterns to form a representative set of frequent behavioral subpaths.

[0095] ;

[0096] in, is the set of frequent behavior subpaths, is the complete set of time series directed behavior path graphs, For the candidate behavioral subpaths, The index of the candidate behavior subpath.

[0097] S4.4. Filter the frequent behavior sub-path set to obtain the sub-path set that deviates from the standard path.

[0098] Furthermore, after obtaining the set of frequent behavior sub-paths, by comparing the structure of each frequent sub-path with the pre-defined standard operation path, analyzing their differences in node sequence, operation logic, time interval, etc., the paths that are consistent with the standard path structure or whose deviations are within the tolerance threshold are screened out, and only sub-paths with significant deviation features such as missing key nodes, sequence dislocation, and abnormal operation delay are retained, finally forming a set of sub-paths that deviate from the standard path.

[0099] S4.5. Perform semantic classification and error labeling on the sub-path set that deviates from the standard path to obtain a set of error pattern templates for the directed behavior path.

[0100] Furthermore, based on the set of sub-paths that deviate from the standard path, relying on the characteristics of node semantics, operation sequence and behavioral logic in each sub-path, combined with the error type dictionary, each sub-path is semantically classified and error-labeled, and sub-paths with similar deviation characteristics are aggregated into corresponding error pattern units. Their node and edge information is further retained in the directed behavioral path structure, and finally a set of directed behavioral path error pattern templates is formed that can be used for behavioral assessment and targeted intervention.

[0101] S5. Match the error pattern template set with the individual behavior path to obtain a scene replay parameter set.

[0102] S5.1. Format the individual behavior path according to the node number and time sequence to obtain the individual behavior path data sequence, encode the node pattern and edge relationship of each error path template in the error pattern template set, and establish an error path template encoding index table.

[0103] Furthermore, each path in the individual behavior path is standardized according to the node numbering sequence and the time sequence of the operations, and uniformly expressed as a time-ordered node sequence to constitute an individual behavior path data sequence; then each error path template in the error pattern template set is structuredly encoded, specifically including assigning a unique node code to each node, assigning directional and sequential attribute labels to the edge relationship between adjacent nodes, thereby forming a path coding string containing the node sequence and edge sequence relationship, further generating a unique index identifier for each coding string, and organizing and establishing an error path template coding index table.

[0104] S5.2. Use the subgraph isomorphism matching algorithm to match the individual behavior path data sequence with the error path template encoding index table to obtain a matching result list.

[0105] Furthermore, in the application of the subgraph isomorphism matching algorithm, the individual behavior path data sequence is regarded as the main graph to be matched, and each encoded path in the error path template encoding index table is regarded as a subgraph to be matched. The node number and time sequence are used as the basic constraints for matching. The subgraph encodings are traversed one by one, and substructures that meet the complete correspondence between node patterns and edge relationships are searched in the main graph. All eligible subgraph isomorphism matches are identified, and all successfully matched subgraph paths and their corresponding matching positions and matching degree information are summarized to generate a matching result list.

[0106] S5.3. Based on the matching result list, trigger conditions and timing information, a set of behavioral deviation fragments is generated, and a preliminary scene replay parameter set is obtained for the deviation type identified in the behavioral deviation fragment set.

[0107] Furthermore, based on the matching result list, combined with the trigger conditions and the timing information of the behavior nodes, the time points and event occurrence sequences of the matching locations are analyzed, and the operation segments that deviate from the standard path are extracted and divided into multiple sets of behavior deviation fragments. For each behavior deviation fragment, the deviation type is identified based on its node characteristics, edge relationships, and time series, such as omitted operations, incorrect sequences, or repeated behaviors. Finally, the identified deviation types and their corresponding behavior fragments are integrated to form a preliminary scenario replay parameter set reflecting the behavior anomaly.

[0108] Specifically, the expression is,

[0109] ;

[0110] in, is the initial scene replay parameter set, For the matching result list, is a set of trigger conditions, It is the timing information.

[0111] S5.4. Dynamically adjust the preliminary scene replay parameter set and the individual behavior path to obtain a scene replay parameter set.

[0112] Furthermore, based on the preliminary scenario replay parameter set, each parameter is dynamically compared and adjusted with the node action and time information in the corresponding individual behavior path. Combined with the changing trend of behavioral deviation and the contextual environment, the specific value and temporal position of the parameter are corrected, and finally a scenario replay parameter set adapted to the individual behavior characteristics is formed to ensure that the parameters can accurately reflect the behavioral deviations and changes in the rehearsal scenario during actual operation.

[0113] S6. Generate a personalized virtual reality emergency process based on the scene replay parameter set.

[0114] S6.1. Analyze the behavior deviation information, trigger conditions, and scene position parameters in the scene replay parameter set to generate an event-driven script configuration table.

[0115] Furthermore, specific behavioral deviation information, corresponding trigger conditions and corresponding scene space position parameters are extracted from the scene replay parameter set. The time node and environmental status of the event are analyzed by combining these data, and the conditions triggering the behavioral deviation are matched with the corresponding positions to generate an event-driven script configuration table containing event triggering timing, triggering conditions, action responses and scene position coordinates.

[0116] S6.2. Based on the event-driven script configuration table, use the three-dimensional scene component library to generate a three-dimensional scene object set.

[0117] Furthermore, based on the event type, trigger conditions and scene position parameters recorded in the event-driven script configuration table, the corresponding models and interactive elements are selected from the 3D scene component library, and the spatial coordinates and time nodes of the event trigger are combined to instantiate and arrange the 3D scene objects one by one to form a complete 3D scene object set that includes dynamic response capabilities and event interaction properties.

[0118] S6.3. Based on the three-dimensional scene object set, establish a drill process timing model.

[0119] Furthermore, based on a collection of 3D scene objects, the system first extracts the event triggering points and their dependencies corresponding to each object, clarifying the sequence and concurrency between events. These events are then arranged chronologically, combined with event triggering conditions and behavioral deviation information to construct a sequential flow chart containing event nodes and directed edges. Next, by defining the execution status and conditional judgments of event nodes, a state transition mechanism for the drill process is formed, ultimately establishing a drill process sequential model that accurately reflects the sequence and interrelationships of events at each stage of the drill.

[0120] Specifically, extract the events associated with each object and their triggering time points from the three-dimensional scene object set, analyze the preconditions and dependencies between events, determine the order and concurrent execution relationships between events, sort the events in chronological order to form a preliminary event sequence, use nodes to represent events, use directed edges to represent the order and conditional dependencies between events, draw an event timing flow chart, define the execution status (not executed, executing, completed) and state transition rules for each event node, combine the trigger conditions to realize automatic advancement of the process, introduce behavioral deviation information and trigger conditions, adjust the process nodes and paths, ensure that the drill process timing model dynamically responds to actual changes in the drill, simulate the timing model, check the logical consistency and completeness, adjust the event relationship and process nodes when necessary, and finally form an accurate and complete drill process timing model.

[0121] S6.4. Obtain the interaction logic mechanism configuration table by practicing the process timing model.

[0122] Furthermore, based on the rehearsal process timing model, the corresponding interactive actions and their triggering conditions for each node are gradually extracted. By combining the temporal sequence and dependency logic between events, the start time and execution order of each action are clarified. At the same time, the correspondence between actions and 3D scene objects is integrated, and finally an interactive logic mechanism configuration table is formed, which includes action triggering conditions, execution order, associated objects, and response feedback.

[0123] Specifically, the expression is,

[0124] ;

[0125] in, Configure the table for the interaction logic mechanism, For the interactive actions, As the trigger condition, For the execution order, In response to feedback.

[0126] S6.5. Organize the three-dimensional scene object set and the interactive logic mechanism configuration table to generate emergency process data.

[0127] Furthermore, by integrating the attributes and status information of each object in the three-dimensional scene object set, and combining it with the trigger conditions, action sequence and response feedback defined in the interactive logic mechanism configuration table, complete emergency process data is constructed; this data clearly describes the behavior and mutual relationship of each scene object in different interactive links, realizes the continuous triggering and response of events in the drill, and ensures the logical coherence and dynamic evolution of the emergency process.

[0128] S6.6. Deploy the emergency process data to the drill platform to generate a personalized virtual reality emergency process.

[0129] Furthermore, the emergency process data is parsed and converted according to the preset drill requirements and individual characteristics, combined with the scene rendering and interaction mechanism in the virtual reality environment, and the corresponding three-dimensional object status and behavior logic are gradually loaded and configured, ultimately generating personalized virtual reality emergency processes customized for the needs of different trainees, achieving a dynamic adaptation and real-time interactive drill experience.

[0130] This embodiment also provides a chemical enterprise intelligent emergency drill system based on virtual reality, including:

[0131] The drill initial data module inputs the multi-dimensional coupling matrix into the virtual reality scene generation unit to obtain the drill initial data;

[0132] The operation trajectory data collection module constructs a behavioral event sequence based on the initial drill data to obtain the operation trajectory data collection;

[0133] The error pattern template module converts the operation trajectory data set into a directed behavior path, identifies the directed behavior path through a graph mining algorithm, and obtains the error pattern template set of the directed behavior path;

[0134] The virtual reality emergency process module matches the error pattern template set with the individual behavior path to obtain the scene replay parameter set, and generates a personalized virtual reality emergency process based on the scene replay parameter set.

[0135] This embodiment also provides a computer device, which is suitable for the case of a virtual reality-based intelligent emergency drill method for chemical enterprises, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the virtual reality-based intelligent emergency drill method for chemical enterprises proposed in the above embodiment.

[0136] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0137] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for implementing a chemical enterprise intelligent emergency drill based on virtual reality as proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0138] In summary, the present invention constructs a multivariable fault evolution knowledge graph and maps it into a multidimensional coupling matrix, thereby achieving structured integration of the causal variables, process flow parameters and accident evolution logic involved in chemical accidents, making the accident causal chain visual and computable in the graph, and providing a semantically rich and logically closed data foundation for the subsequent virtual rehearsal scene generation, thereby improving the pertinence and complexity expression ability of the rehearsal content. By matching the error pattern template set with the individual behavior path, a scene replay parameter set is obtained, which realizes the automatic identification and parameterized extraction of behavioral deviations and error types, and dynamically adjusts them in combination with individual paths, thereby driving the personalized replay configuration of the virtual scene, effectively enhancing the behavior-driven mechanism and emergency feedback adaptability of the rehearsal process.

[0139] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A virtual reality-based intelligent emergency drill method for chemical enterprises, characterized by: include, Collect individual behavior paths, historical accident data, and chemical production process parameters. Based on these data, a multivariate fault evolution knowledge graph is constructed to obtain a multidimensional coupling matrix. Input the multi-dimensional coupling matrix into the virtual reality scene generation unit to obtain the initial data of the drill; Based on the initial data of the drill, a behavioral event sequence is constructed to obtain a set of operation trajectory data; The operation trajectory data set is converted into a directed behavior path, and the directed behavior path is identified through a graph mining algorithm to obtain a set of error pattern templates of the directed behavior path; Match the error pattern template set with the individual behavior path to obtain the scene replay parameter set, and generate a personalized virtual reality emergency process based on the scene replay parameter set; The operation trajectory data set is converted into a directed behavior path, and the directed behavior path is identified by a graph mining algorithm to obtain a set of error pattern templates of the directed behavior path, including the following steps: Extract action nodes and time sequence relationships from the operation trajectory data set and construct the initial directed graph of operation events; Sort the initial directed graph of operation events by the operation time sequence to obtain a time series directed behavior path graph; Based on the frequent subgraph mining algorithm, the pattern of the time series directed behavior path graph is extracted to identify the frequent behavior subpath set; Filter the frequent behavior sub-path set to obtain the sub-path set that deviates from the standard path; The set of subpaths that deviate from the standard path are semantically classified and annotated with error labels to obtain a set of error pattern templates for directed behavior paths.

2. The virtual reality-based intelligent emergency drill method for chemical enterprises according to claim 1, characterized in that: Collect individual behavior paths, historical accident data and chemical production process parameters, build a multivariate fault evolution knowledge graph based on historical accident data and chemical production process parameters, and obtain a multidimensional coupling matrix, including the following steps: Collect the operational behavior sequences and response decision data generated by trainees in different emergency drill tasks to form individual behavior paths; Organize the chemical enterprise accident case database and extract chemical production process parameters from the equipment layout of chemical enterprises; Utilize graph structure modeling methods to construct a multivariate fault evolution knowledge graph from historical accident data and chemical production process parameters; Based on the multivariable fault evolution knowledge graph, it is mapped into a multidimensional coupling matrix.

3. The virtual reality-based intelligent emergency drill method for chemical enterprises according to claim 2, characterized in that: Input the multi-dimensional coupling matrix into the virtual reality scene generation unit to obtain the initial data of the drill. The following steps are included: The accident cause combinations and variable evolution path information are extracted from the multidimensional coupling matrix. Based on the accident cause combinations, the spatial coordinates of the fault source are located in the three-dimensional equipment model. Based on the variable evolution path information, the diffusion range of hazardous substances is obtained through physical modeling. Based on the diffusion range of dangerous substances and individual behavior paths, the evacuation path point set is obtained; According to the variable evolution path information, set the animation response logic for the 3D model; The diffusion range of hazardous materials, evacuation route point set, animation response logic and fault source spatial coordinates are input into the virtual reality scene to generate initial data for the drill.

4. The virtual reality-based intelligent emergency drill method for chemical enterprises according to claim 3, characterized in that: Based on the initial data of the drill, a behavioral event sequence is constructed to obtain an operation trajectory data set, which includes the following steps: Read the initial data of the drill and obtain the set of interactive objects; Perform task mapping on the interactive object set, generate a standard operation event list, record the trainees' operations in the virtual reality environment, and obtain the original behavior action flow; Perform event semantic annotation on the standard operation event list and the original behavior action flow to form a preliminary behavior event sequence; The preliminary behavior event sequence is subjected to trajectory extraction to obtain a set of behavior trajectory fragments, which are then aggregated by individuals to obtain an operation trajectory data set.

5. The virtual reality-based intelligent emergency drill method for chemical enterprises according to claim 1, characterized in that: Matching the error pattern template set with the individual behavior path to obtain the scene replay parameter set includes the following steps: Format the individual behavior path according to the node number and time sequence to obtain the individual behavior path data sequence, encode the node pattern and edge relationship of each error path template in the error pattern template set, and establish the error path template encoding index table; Use the subgraph isomorphism matching algorithm to match the individual behavior path data sequence with the error path template encoding index table to obtain a matching result list; Based on the matching result list, trigger conditions and timing information, a set of behavioral deviation fragments is generated, and the deviation type identified in the set of behavioral deviation fragments is used to obtain a preliminary scene replay parameter set; The preliminary scene replay parameter set and the individual behavior path are dynamically adjusted to obtain the scene replay parameter set.

6. The virtual reality-based intelligent emergency drill method for chemical enterprises according to claim 5, characterized in that: Generate personalized virtual reality emergency procedures based on the scene replay parameter set, The following steps are included: Parse the behavior deviation information, trigger conditions and scene position parameters in the scene replay parameter set to generate an event-driven script configuration table; Based on the event-driven script configuration table, a 3D scene object set is generated using the 3D scene component library; Based on the 3D scene object set, a drill process timing model is established; By rehearsing the process sequence model, we can obtain the configuration table of the interaction logic mechanism; Organize the 3D scene object collection and interactive logic mechanism configuration table to generate emergency process data; Deploy emergency process data to the drill platform to generate personalized virtual reality emergency processes.

7. A virtual reality-based intelligent emergency drill system for chemical enterprises, based on the virtual reality-based intelligent emergency drill method for chemical enterprises according to any one of claims 1 to 6, characterized in that: include, The multi-dimensional coupling matrix module collects individual behavior paths, historical accident data, and chemical production process parameters. Based on the historical accident data and chemical production process parameters, it constructs a multi-variable fault evolution knowledge graph to obtain a multi-dimensional coupling matrix. The drill initial data module inputs the multi-dimensional coupling matrix into the virtual reality scene generation unit to obtain the drill initial data; The operation trajectory data collection module constructs a behavioral event sequence based on the initial drill data to obtain the operation trajectory data collection; The error pattern template module converts the operation trajectory data set into a directed behavior path, identifies the directed behavior path through a graph mining algorithm, and obtains the error pattern template set of the directed behavior path; The virtual reality emergency process module matches the error pattern template set with the individual behavior path to obtain the scene replay parameter set, and generates a personalized virtual reality emergency process based on the scene replay parameter set.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the chemical enterprise intelligent emergency drill method based on virtual reality are implemented as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the virtual reality-based intelligent emergency drill method for chemical enterprises according to any one of claims 1 to 6 are implemented.

Images