Emergency drill simulation method and system using war game

Abstract

The application relates to the technical field of emergency, in particular to an emergency drill simulation method and system applying war game deduction, wherein the method comprises the following steps: obtaining a HAZOP report, and analyzing risk element data in the HAZOP report; constructing a risk evolution path atlas based on the risk element data, storing the risk evolution path atlas as structured data, and obtaining a structured risk data set; mapping chess pieces based on the structured risk data set, constructing a war game deduction model based on a preset evolution rule to perform deduction simulation, and outputting a war game deduction time series data set. By introducing the war game deduction mechanism in the emergency drill simulation, the whole-process digitization from historical risk information to the simulation deduction process can be realized, so that the decision maker can be assisted in risk prediction and response strategy evaluation.

Description

Technical Field

[0001] This application relates to the field of emergency technology, and in particular to emergency drill simulation methods and systems using wargaming. Background Technology

[0002] With the continuous deepening of emergency management informatization, existing emergency drill platforms have made significant progress in drill simulation technology. Early drill systems mainly relied on static contingency plans and simple process simulations. In recent years, they have gradually developed into comprehensive platforms integrating multiple technologies such as Geographic Information Systems (GIS), 3D visualization, and discrete event simulation. Some advanced systems have also introduced virtual reality (VR) and augmented reality (AR) technologies, achieving highly immersive scene reconstruction and interactive training. At the same time, some platforms have begun to combine system dynamics models with data-driven methods to evaluate the dispatch efficiency of emergency resources and overall response capabilities, promoting the transformation of drills from "process demonstrations" to "realistic simulations."

[0003] However, current technologies still have some significant shortcomings in application. First, most systems rely on pre-set scripts, lacking the ability to flexibly respond to emergencies and failing to realistically reflect the uncertainties and dynamic changes in complex environments. Second, existing simulations are mostly concentrated at the strategic or operational level, with relatively crude depictions of key details such as tactical execution, on-site command, and multi-departmental collaboration, leading to discrepancies between simulation results and actual operations. Furthermore, systems generally lack effective decision-making feedback mechanisms, failing to provide real-time evaluation and guidance for participants' judgments and actions, thus affecting the learning outcomes of exercises. Simultaneously, the data integration between modules is low, making it difficult to form a unified and coherent simulation environment, limiting the overall system's practicality and scalability. Summary of the Invention

[0004] Firstly, this application provides a simulation method for emergency drills using wargaming, including: The report parsing steps involve obtaining historical risk analysis reports and analyzing the risk element data in the HAZOP report (Hazard and Operability Study, or HAZOP for short); The dataset acquisition step involves constructing a risk evolution path map based on the risk element data, storing the risk evolution path map as structured data, and obtaining a structured risk dataset. The structured risk dataset is used to clarify the relationship between accident sources, derivative paths, and influencing factors. The structured risk dataset includes accident ID field, triggering condition field, triggering time field, derivative consequences field, evolution probability field, impact range field, and quantitative attribute parameter field. The simulation process involves mapping chess pieces based on the structured risk dataset, constructing a wargame simulation model based on preset evolution rules, generating preliminary scenarios and performing dynamic simulations, and outputting a wargame simulation time series dataset to achieve the prediction and assessment of emergency risks.

[0005] Based on the above steps, this application introduces a wargaming mechanism into emergency drill simulations, enabling full-process digitization from historical risk information to the simulation process. By clarifying the logical connections between accident sources, evolution paths, and influencing factors through a structured risk dataset, it helps establish a systematic risk evolution model. Compared to traditional static drills, this method can dynamically present accident development trends, improving the realism and targeted response of the drills. Simultaneously, by quantifying parameter control and setting evolution probabilities, the scientific rigor and accuracy of the simulation are enhanced. The resulting wargaming time-series dataset not only assists decision-makers in risk prediction and response strategy evaluation but also possesses strong versatility and scalability, making it suitable for simulating different types of emergency scenarios.

[0006] In some embodiments, the report parsing step includes: The text semantic parsing step extracts risk element data from the HAZOP report based on a pre-trained text parsing model. The risk element data includes process nodes, risk sources, trigger prompts, trigger conditions, consequence chains, and safety barriers. The consequence chains include initial accidents, derivative accidents, and secondary hazards. The text parsing model is trained on a professional HAZOP corpus and has the ability to extract semantic-level information. It can identify key risk element data in the report, including process nodes, risk sources, trigger prompts, trigger conditions, consequence chains, and corresponding safety barriers. The extracted results are stored in a structured manner for subsequent analysis. The accident analysis step involves determining the probability of occurrence of the initial accident, derivative accident, and secondary hazard in the consequence chain based on the triggering conditions and consequence chain in the risk element data, and predicting the occurrence delay and impact range of the accident in the consequence chain based on the consequence quantification analysis model, so as to clarify the regional boundaries corresponding to different consequences. The impact range is constrained by time and occurrence probability.

[0007] In the accident analysis step, this application first establishes an accident evolution model based on the extracted triggering conditions and consequence chains. Using historical data or expert systems, combined with probabilistic models such as Bayesian networks and fault trees, the probability of occurrence at each level of the accident is determined. Simultaneously, a consequence quantification analysis model is introduced to predict the occurrence delay and impact range of various types of accidents based on accident type and process conditions. The impact range is further constrained by the coupling of accident occurrence probability and evolution time, accurately depicting the boundary distribution of accident development in spatial and temporal dimensions, thereby providing precise input parameters and spatial scenario setting basis for subsequent wargaming simulations.

[0008] Based on the aforementioned text semantic parsing and accident analysis steps, a dual mechanism of semantic parsing and quantitative analysis is introduced into the report parsing process. This not only improves the automation and accuracy of risk factor identification but also enables precise modeling of the dynamic evolution of accidents through quantitative models. By clarifying the probability and time response characteristics of various accidents in the consequence chain, the scientific nature and data support capabilities of accident assessment are significantly enhanced, providing a reliable basis for the construction of emergency simulation scenarios. Simultaneously, the improved accuracy of regional boundary determination makes the simulation scenarios closer to actual accident situations, helping to optimize emergency response strategies and resource allocation, and enhancing the practical value of exercises and the relevance of strategy formulation.

[0009] In some embodiments, the text parsing model is fine-tuned based on a BERT (Bidirectional Encoder Representations from Transformers) model pre-trained. The BERT model, through large-scale pre-training on general corpora, possesses a deep understanding of complex semantic relationships. Subsequently, for the specific application scenario of HAZOP risk analysis reports, the BERT model undergoes supervised fine-tuning training. A dedicated dataset is constructed using labeled HAZOP corpora to enhance its ability to identify technical terms and risk element structures. During the fine-tuning process, tasks such as classification and sequence labeling guide the model to learn and identify target entities such as process nodes, risk sources, trigger prompts, trigger conditions, consequence chains, and safety barriers, and output structured risk element data. This structured output further provides high-quality data input for subsequent accident analysis and simulation steps, improving the overall system's intelligence level and semantic parsing accuracy.

[0010] In some embodiments, the dataset acquisition step further includes: The risk evolution path graph definition steps involve creating entity nodes and relationship edges based on the risk element data, mapping and generating a risk evolution path graph. The entity nodes in the risk evolution path graph include: risk source, triggering event, accident consequence, accident impact, and emergency resources. The relationship edges include: triggering relationship, derivative relationship, dependency relationship, and impact relationship, so as to dynamically display the evolution path of the initial accident, derivative accident, and secondary hazard. The relationship attributes are configured for entity relationships in combination with the occurrence delay and impact range. The dataset generation step transforms the risk evolution path map into structured data to obtain a structured risk dataset. The structured risk dataset includes fields such as accident ID, triggering condition, triggering time, accident consequence, evolution probability, affected object, affected scope, and quantitative attribute parameter.

[0011] Based on the above steps, this application models the risk elements extracted from the HAZOP report using a graph structure, making the accident evolution logic more intuitive and clear. In the graph definition step, the system creates multiple types of entity nodes based on risk semantic information, including risk sources, triggering events, accident consequences, accident impacts, and emergency resources. Simultaneously, it defines various semantic relationship edges, such as triggering relationships, derivative relationships, dependency relationships, and impact relationships. These relationship edges are further combined with the accident's occurrence delay and spatial impact range, and additional configuration relationship attributes are added to represent information such as the time span, impact intensity, and geographical boundaries during the evolution process, thereby constructing a graph model that can dynamically display the accident evolution path.

[0012] Subsequently, the aforementioned graph structure is transformed into a structured risk dataset. The transformation process includes steps such as extracting information from nodes and edges, standardizing attributes, and classifying fields. The final dataset uses a field-based format to represent event characteristics, with fields including event ID, triggering conditions, triggering time, event consequences, evolution probability, affected objects, scope of impact, and other quantitative attribute parameters. This structured data can be stored and transmitted in JSON format, exhibiting good cross-platform compatibility and parsing efficiency, facilitating subsequent use in wargaming models and graphical visualization.

[0013] This application introduces a graph definition mechanism to achieve visualized modeling and dynamic evolution representation of risk information. This not only enhances the expressive power of risk data but also improves the accuracy and completeness of accident evolution path modeling. The classification design of entity nodes and relation edges allows for the differentiated management of different types of risk information, supporting more complex logical relationship modeling. The introduction of relation attributes enables the graph to have dynamic display capabilities, presenting the risk propagation process based on the temporal and spatial characteristics of accident development, improving the timeliness and relevance of risk prediction and response strategy formulation. Simultaneously, by transforming the graph structure into a standardized field dataset, a pathway between knowledge modeling and numerical analysis is established, providing a solid data foundation for subsequent intelligent inference and strategy generation. Using JSON format for storage enhances the system's scalability and data exchange efficiency, facilitating integration with multi-source platforms.

[0014] In some embodiments, the accident analysis step further includes: The failure rate of the risk source corresponding to the process node is obtained as the failure rate of the base event, and the data on the operator error rate and the historical fluctuation data of the process parameters are obtained as the failure rate of the intermediate event. The failure rate data is obtained through the Offshore Reliability Data (OREDA) database, the reliability data provided by the equipment manufacturer, and the equipment failure records of the industry / enterprise. The operator error rate data is determined through the industry operation error statistical report and the enterprise safety training and assessment records. The historical fluctuation data of the process parameters includes excessively high outlet pressure, differential pressure indication of inlet and outlet materials, and remote transmission of the outlet flow of the dry benzene circulation pump. Specifically, the frequency of key parameters (such as pressure and flow) exceeding the normal range can be calculated from the DCS / SCADA historical database as the probability of the "intermediate event". The initial accident is constructed based on the Fault Tree Analysis (FTA) model. The intermediate and bottom events associated with the initial accident are entered. The failure rate of the intermediate events is obtained by performing an OR operation based on the failure rate of the bottom events. The annual occurrence probability of the initial accident is obtained by performing an OR operation based on the failure rate of the intermediate events. Identify the derivative accidents and / or secondary hazards associated with the initial accident, configure trigger probabilities for the derivative accidents and / or secondary hazards based on industry experience or expert judgment, and calculate the annual occurrence probability of the derivative accidents and / or secondary hazards based on the annual occurrence probability of the initial accident and the trigger probabilities.

[0015] Based on the above steps, a fault tree (FTA) model is introduced to quantitatively model various risk events in the accident evolution process. Risk sources associated with process nodes are used as base events, and failure rate data are obtained from multiple data sources, including the OREDA (Offshore Reliability Data) database, reliability manuals provided by equipment manufacturers, and industry or enterprise equipment operation and maintenance records. Subsequently, key influencing factors at the operational and process levels are extracted as intermediate events; the initial accident is used as the top event, and its corresponding intermediate and base events are identified. An OR gate is used to synthesize the base events to obtain the failure rate of the intermediate events; then, an OR gate is used to synthesize the intermediate events again to obtain the annual probability of the top event (i.e., the initial accident). Accident risk analysis shifts from qualitative speculation to quantitative calculation, improving the scientific rigor and reliability of risk modeling. The FTA model allows for systematic analysis of the accident causal chain, clarifying the failure contribution of equipment, personnel, and process parameters at each level, facilitating the precise formulation of control strategies. The wide range and traceability of data sources enhance the model's practical operability. Furthermore, by calculating the annual probability of occurrence in a hierarchical manner, the simulation model is equipped with quantitative support, thereby improving the credibility and prediction accuracy of emergency drills.

[0016] In some embodiments, the accident analysis step further includes: The occurrence delay and impact range of the initial and derivative accidents are predicted based on the consequence quantification analysis model. The consequence quantification analysis model includes at least the following: national standard calculation model, Gaussian diffusion model / heavy gas diffusion model SLAB, pool fire thermal radiation model, vapor cloud explosion thermal radiation model, jet fire calculation model, explosion model, etc.

[0017] Using the above model, this application can combine different models according to the accident scenario type to predict the damage diffusion process of the initial accident and its derivative events in the spatial and temporal dimensions, which constitutes an important input basis for the wargaming simulation model.

[0018] This implementation method introduces multiple engineering consequence analysis models to achieve scientific prediction of accident consequences, especially dynamic modeling of the impact range and response delay, thereby improving the accuracy and reliability of accident simulation. Different models are adapted to different types of accidents, providing technical support for multi-scenario and multi-source risk modeling; the input parameters are readily available and the output results have engineering interpretability, facilitating the formation of spatialized and quantitative evolutionary scenarios by combining with graphical models, thus improving the data support capability and scenario reproduction of wargaming simulations.

[0019] In some embodiments, the simulation step includes: The steps for defining wargame pieces include configuring emergency drill entities, accident scenarios, and accident sites, and constructing and defining wargame pieces and their attributes based on the structured risk dataset. The wargame pieces include entity pieces, accident pieces, and board pieces. Piece attributes include type, position, status, radius of influence, operation rules, and controlled range.

[0020] The simulation rule construction steps involve inputting the accident type, environmental variables, and emergency resource configuration into the rule parameter configurator to generate a baseline simulation rule library. This library includes entity attribute evolution rules, accident evolution rules, and adversarial rules. Entity attribute evolution rules describe the changes in the ability attributes of entity pieces, such as changes in ability attributes over time, on the chessboard (terrain, environment, resources), and the impact of accident pieces. Accident evolution rules, based on accident maps and consequence analysis models, specify how the accident situation evolves over time and how pieces change, such as a leaking piece becoming a fire piece. Adversarial rules define the interaction mechanisms, response logic, and effect evaluation standards between pieces, including joint adversarial interactions and dynamic mutual influence among entity pieces, accident pieces, and on the chessboard. The adversarial exercise steps, based on the benchmark simulation rule base, simulate the evolution of accidents and emergency confrontation processes through a round-based, real-time event-driven hybrid simulation algorithm and a random event generator, and output the wargame simulation time series dataset.

[0021] In one embodiment, the deduction algorithm uses N minutes as the deduction step size, and each round of deduction is processed in the following order: automatically advance and update the state of the accident piece according to the deduction rules; receive the action instructions of the physical piece and execute the effect; determine whether the triggering condition is met to generate a new accident piece; apply a random event generator to introduce sudden events to affect the current confrontation strategy; record the state of all pieces, response actions, and accident situation in the current round, and write them into the wargame deduction time series dataset.

[0022] The simulation can continue until all incident situations stabilize or resources are exhausted, at which point the system will finally output complete data on the confrontation exercise process, including evolution paths, response behaviors, and effectiveness evaluation indicators.

[0023] By employing wargaming modeling and turn-based drills, this implementation transforms traditional static emergency drills into an interactive and evaluable dynamic simulation system, enhancing the understanding and response capabilities to the evolution of complex accidents. The introduction of a dual definition mechanism for pieces and rules effectively organizes various resources and scenario elements, improving modeling flexibility. The integration of a random event mechanism enhances the handling of uncertainties in the simulation, making the drills more closely resemble real-world scenarios. Simultaneously, the system can output standardized time-series datasets, facilitating subsequent evaluation and post-mortem analysis, and providing data support for emergency management decision-making.

[0024] The entity chess pieces include pieces used to represent corresponding emergency drill entities. Emergency drill entities include rescue and disposal personnel, vehicles, and disposal equipment. The entity chess pieces are configured with basic attributes, positioning attributes, capability parameters, and exclusive permissions. The basic attributes include role type and survival value. The positioning attributes include coordinates and movement speed. The entity chess piece's capability parameters include fire extinguishing capability, casualty rescue capability, and material transportation capability. The coordinates include the initial position and current coordinates. The fire extinguishing capability includes fire extinguishing efficiency value, effective range, and duration. The casualty rescue capability includes rescue capacity and rescue radius. The material transportation capability includes carrying capacity and support range.

[0025] For ease of understanding, the following examples of physical chess pieces are provided in this application: Piece: "Water Tanker Fire Truck", Attributes: Coordinates (X,Y), Speed ​​3m / s, Fire Extinguishing Efficiency 40L / s, Range 65m, Continuous Spraying for 20 Minutes; Piece: "Medical Team", Attributes: Rescue capacity 2 people / time, response time 5 minutes, coverage radius 100m.

[0026] The accident pieces are constructed based on the structured risk dataset and represent accident nodes and evolution states, such as leak sources, ignition points, and explosion centers. Each accident piece is configured with a state switching attribute, such as switching from leak to combustion, and from combustion to explosion. Optionally, the accident pieces' attributes also include: accident piece name / accident ID, accident center point coordinates, type, state (triggered - not triggered), triggering condition, triggering probability, spread rate, derivative consequences (converted to other accident pieces), hazard level, and quantitative attribute parameters. A passage speed penalty coefficient is determined based on the influence range. This coefficient indicates that when a passage path passes through the influence range, the passage speed is inversely proportional to the distance between the passage point and the accident center point. The accident piece subsequently appears or disappears in a consequence chain.

[0027] The chessboard and chess pieces include controlled areas, safe areas, and resource areas of the training ground. Each chessboard and chess piece is configured with attributes. Controlled areas are areas where passage is prohibited or damage is caused, such as fire zones. Their attributes include: coordinate range, passage penalty value, damage value, and time window. Safe areas allow for evacuation or centralized evacuation of personnel. Their attributes include: coordinate range, protection level, open duration, and maximum capacity. Resource areas provide emergency resources or capability enhancements, such as emergency supply depots, fire hydrants, fire monitors, and fire dikes. Their attributes include: resource type (supplies, water), enhancement value, and target.

[0028] Based on the above configuration, this application embodiment constructs a mapping table between HAZOP risk factors and wargame piece attributes through multi-dimensional attribute mapping, realizing dynamic attribute configuration: for example, the leakage rate is set to 0.5~2m³ / h and can be adjusted through a slider interface; the capabilities of physical pieces and the impact range of accident pieces are interactively calculated to form an actual operable space; collision judgment, influence effect and status linkage are supported between different pieces.

[0029] In some embodiments, the preset rule parameter configurator of this application automatically generates a deduction rule library with logical conditions and action results based on input parameters such as accident type, exercise environment and resource configuration, for the deduction engine to call.

[0030] In another embodiment, the rule parameter configurator can parse the input dataset and configuration items, and extract key variables, such as... Accident characteristic variables: accident type, triggering conditions, evolution path, evolution probability, scope of impact, etc.; Environmental variables: meteorological information (wind speed, wind direction, temperature), spatial structure (regional boundaries, distribution of resource points); Resource configuration parameters: rescue team type, number, initial coordinates, response capability, speed, etc.; Scene status indicators: current status of various chess pieces, available actions, skill trigger conditions, etc.

[0031] The preset rule parameter configurator of this application has multiple built-in rule template sets. The parameter configurator calls the corresponding template according to the extracted variables, fills the parameter fields and generates rule statements to form a structured rule object.

[0032] Then, the rule parameter configurator models the rules using a cause-effect graph, connecting triggering events, conditional decisions, and result states into a state transition graph. This automatically detects logical loops, prevents infinite loops or triggering conflicts, and supports parallel rule settings with multi-path triggering.

[0033] The rule parameter configurator uses structured input parsing, template matching, condition-action modeling, and random mechanism injection to form a complete set of inference rules covering the entire process of accident development and resource response behavior. This rule base provides the wargaming system with the basis for action and a state control framework, realizing a logical closed loop from structured risk data to dynamic exercise behavior, ensuring that the exercise simulation has a clear causal chain, complete decision-making conditions, and controllable emergency paths.

[0034] In some embodiments, the random event generator uses an improved Monte Carlo algorithm to introduce the probability distribution of historical drill accidents to generate sudden interference events.

[0035] Generating random perturbation events through Monte Carlo simulation not only enhances the dynamic complexity and realism of wargames but also systematically verifies the stability and response effectiveness of contingency strategies under changing conditions. This mechanism combines flexibility and controllability, allowing for adjustments to the intensity and frequency of perturbations based on the exercise objectives, while also providing quantitative data support for subsequent risk assessment and algorithm optimization.

[0036] In some embodiments, the hybrid inference algorithm performs the following process in each round: Update the status and impact range of the accident chess piece (based on the diffusion model); Execute entity chess actions (such as moving, extinguishing fires, and rescuing) according to the rule base; Call the random event generator to determine if a sudden disturbance has occurred; Determine whether the accident status has changed (e.g., cessation of evolution, state transition); Write the timing data and advance to the next round. That is, after each round of simulation, output a standardized timing data.

[0037] In some embodiments, each time series data point in the wargaming simulation time series dataset includes: The event includes a timestamp, event type, scope of impact, list of affected pieces, and list of unaffected pieces. The timestamp includes the number of rounds and the cumulative time.

[0038] This implementation combines turn-based and spatial diffusion calculations to achieve dynamic simulation of accident situations and refined tracking of entity response behaviors, enhancing the realism, data-driven nature, and assessability of emergency drills. Through standardized temporal data structures and spatial coordinate systems, a two-way mapping bridge is established between the wargaming system and real-world scenarios, facilitating the reuse of drill scenarios, data mining, and strategy verification.

[0039] Secondly, embodiments of this application provide an emergency drill simulation system that applies wargaming, including: The report parsing unit is used to obtain the HAZOP report and parse the risk element data in the HAZOP report; The dataset acquisition unit is used to construct a risk evolution path map based on the risk element data, store the risk evolution path map as structured data, and obtain a structured risk dataset. The structured risk dataset is used to clarify the relationship between accident sources, derivative paths, and influencing factors. The structured risk dataset includes accident ID field, triggering condition field, triggering time field, derivative consequences field, evolution probability field, impact range field, and quantitative attribute parameter field. The simulation unit is used to map chess pieces based on the structured risk dataset, construct a wargame simulation model based on preset evolution rules, generate a preliminary scenario and perform dynamic simulation, and output a wargame simulation time series dataset.

[0040] Thirdly, embodiments of this application provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the emergency drill simulation method for applying wargaming as described in the first aspect above.

[0041] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0042] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart illustrating the emergency drill simulation method according to an embodiment of this application; Figure 2 This is another flowchart illustrating the emergency drill simulation method according to an embodiment of this application; Figure 3 This is another flowchart illustrating the emergency drill simulation method according to an embodiment of this application; Figure 4 This is another flowchart illustrating the emergency drill simulation method according to an embodiment of this application; Figure 5 This is a structural block diagram of an emergency drill simulation system according to an embodiment of this application. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0044] Obviously, the accompanying drawings described below are merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar scenarios based on these drawings without any inventive effort. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, any changes to design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as insufficient disclosure of the content of this application.

[0045] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0046] This embodiment provides a simulation method for emergency drills using wargaming. Figures 1 to 4 This is a flowchart of an emergency drill simulation method using wargaming according to an embodiment of this application, such as... Figures 1 to 4 As shown, the process includes the following steps: Report parsing step S1: Obtain the HAZOP report and parse the risk element data in the HAZOP report; In the dataset acquisition step S2, a risk evolution path map is constructed based on the risk element data, and the risk evolution path map is stored as structured data to obtain a structured risk dataset. The structured risk dataset is used to clarify the relationship between accident sources, derivative paths, and influencing factors. The structured risk dataset includes accident ID field, triggering condition field, triggering time field, derivative consequences field, evolution probability field, impact range field, and quantitative attribute parameter field. In simulation step S3, chess pieces are mapped based on the structured risk dataset, and a wargaming simulation model is constructed based on preset evolution rules to generate a preliminary scenario and perform dynamic simulation. The wargaming simulation time series dataset is then output to achieve the prediction and assessment of emergency risks.

[0047] Based on the above steps, this application introduces a wargaming mechanism into emergency drill simulations, enabling full-process digitization from historical risk information to the simulation process. By clarifying the logical connections between accident sources, evolution paths, and influencing factors through a structured risk dataset, it helps establish a systematic risk evolution model. Compared to traditional static drills, this method can dynamically present accident development trends, improving the realism and targeted response of the drills. Simultaneously, by quantifying parameter control and setting evolution probabilities, the scientific rigor and accuracy of the simulation are enhanced. The resulting wargaming time-series dataset not only assists decision-makers in risk prediction and response strategy evaluation but also possesses strong versatility and scalability, making it suitable for simulating different types of emergency scenarios.

[0048] In some embodiments, the report parsing step S1 includes: In text semantic parsing step S11, risk element data from the HAZOP report is extracted based on a pre-trained text parsing model. This risk element data includes process nodes, risk sources, triggering prompts, triggering conditions, consequence chains, and safety barriers. The consequence chain includes initial accidents, derivative accidents, and secondary hazards. The text parsing model is trained on a professional HAZOP corpus and possesses the ability to extract semantic-level information. It can identify key risk element data in the report, including process nodes (such as reactors and pipelines), risk sources (such as excessive pressure and flammable material leaks), triggering prompts (such as alarm signals and operational anomalies), triggering conditions (such as temperature thresholds and equipment malfunctions), consequence chains (including initial accidents, derivative accidents, and secondary hazards), and corresponding safety barriers (such as automatic shut-off systems and safety valves). The extracted results are stored in a structured manner for subsequent analysis. In the accident analysis step S12, the probability of occurrence of the initial accident, derivative accident and secondary hazard in the consequence chain is determined based on the triggering conditions and consequence chain in the risk element data. Based on the consequence quantification analysis model, the occurrence delay and impact range of the accident in the consequence chain are predicted to clarify the regional boundaries corresponding to different consequences. The impact range is constrained by time and occurrence probability.

[0049] In the accident analysis step S12, this application first establishes an accident evolution model based on the extracted triggering conditions and consequence chains. Using historical data or expert systems, combined with probabilistic models such as Bayesian networks and fault trees, the probability of occurrence of each level of accident (initial accident, derivative accident, secondary hazard) is determined. Simultaneously, a consequence quantification analysis model (such as a CFD simulation model or a shock wave propagation model) is introduced to predict the occurrence delay (i.e., the time delay from the triggering of the initial triggering condition to the occurrence of the initial accident) and its impact range based on the accident type and process conditions. The impact range is further constrained by the coupling of the accident occurrence probability and evolution time, accurately depicting the boundary distribution of accident development in the spatial and temporal dimensions, thereby providing precise input parameters and spatial scenario setting basis for subsequent wargaming simulations.

[0050] Based on steps S11 and S12 above, a dual mechanism of semantic parsing and quantitative analysis is introduced into the report parsing process. This not only improves the automation and accuracy of risk factor identification but also enables precise modeling of the dynamics of accident evolution through quantitative models. By clarifying the probability and time response characteristics of various accidents in the consequence chain, the scientific nature and data support capabilities of accident assessment are significantly enhanced, providing a reliable basis for the construction of emergency simulation scenarios. Simultaneously, the improved accuracy of regional boundary determination makes the simulation scenarios closer to actual accident situations, helping to optimize emergency response strategies and resource allocation, and improving the practical value of exercises and the relevance of strategy formulation.

[0051] In some embodiments, the text parsing model is fine-tuned based on pre-training of the BERT model. The BERT model, through large-scale pre-training on general corpora, possesses a deep understanding of complex semantic relationships. Subsequently, for the specific application scenario of HAZOP risk analysis reports, the BERT model undergoes supervised fine-tuning training. A dedicated dataset is constructed using labeled HAZOP corpora to enhance its ability to identify technical terms and risk element structures. During the fine-tuning process, tasks such as classification and sequence labeling guide the model to learn and identify target entities such as process nodes, risk sources, trigger prompts, trigger conditions, consequence chains, and safety barriers, and output structured risk element data. This structured output further provides high-quality data input for subsequent accident analysis and simulation steps, improving the overall system's intelligence level and semantic parsing accuracy.

[0052] In some embodiments, the dataset acquisition step S2 further includes: In the graph definition step S21, entity nodes and relationship edges are created based on the risk element data, and a risk evolution path graph is generated. The entity nodes in the risk evolution path graph include: risk source, triggering event, accident consequence, accident impact, and emergency resources. The relationship edges include: triggering relationship, derivative relationship, dependency relationship, and impact relationship, so as to dynamically display the evolution path of the initial accident, derivative accident, and secondary hazard. The relationship attributes are configured for the entity relationships in combination with the occurrence delay and impact range. In the dataset generation step S22, the risk evolution path map is converted into structured data to obtain a structured risk dataset. The structured risk dataset includes accident ID field, trigger condition field, trigger time field, accident consequence field, evolution probability field, affected object field, affected scope field, and quantitative attribute parameter field. Optionally, the structured data can be in JSON format or XML format, etc.

[0053] Based on the above steps, this application models the risk elements extracted from the HAZOP report using a graph structure, making the accident evolution logic more intuitive and clear. In the graph definition step S21, the system creates multiple types of entity nodes based on risk semantic information, including risk sources (such as high-voltage equipment), triggering events (such as temperature rise), accident consequences (such as explosion), accident impacts (such as casualties and property damage), and emergency resources. Emergency resources include rescue resources and monitoring resources. Rescue resources may include medical teams, fire brigades, and rescue supplies / equipment facilities, while monitoring resources may include monitoring and alarm devices. Simultaneously, various semantic relationship edges are defined, such as triggering relationships (representing causal relationships between events), derivative relationships (representing the hierarchical development of accident consequences), dependency relationships (describing resource calls and event dependencies), and impact relationships (representing the impact of accident consequences on external objects). These relationship edges are further combined with the time delay and spatial impact range of the accident, and additional configuration relationship attributes are added to represent information such as the time span, impact intensity, and geographical boundaries during the evolution process, thereby constructing a graph model that can dynamically display the accident evolution path.

[0054] Subsequently, the aforementioned graph structure is transformed into a structured risk dataset. The transformation process includes steps such as extracting information from nodes and edges, standardizing attributes, and classifying fields. The final dataset uses a field-based format to represent event characteristics, with fields including event ID, triggering conditions, triggering time, event consequences, evolution probability, affected objects, scope of impact, and other quantitative attribute parameters. This structured data can be stored and transmitted in JSON format, exhibiting good cross-platform compatibility and parsing efficiency, facilitating subsequent use in wargaming models and graphical visualization.

[0055] This application introduces a graph definition mechanism to achieve visualized modeling and dynamic evolution representation of risk information. This not only enhances the expressive power of risk data but also improves the accuracy and completeness of accident evolution path modeling. The classification design of entity nodes and relation edges allows for the differentiated management of different types of risk information, supporting more complex logical relationship modeling. The introduction of relation attributes enables the graph to have dynamic display capabilities, presenting the risk propagation process based on the temporal and spatial characteristics of accident development, improving the timeliness and relevance of risk prediction and response strategy formulation. Simultaneously, by transforming the graph structure into a standardized field dataset, a pathway between knowledge modeling and numerical analysis is established, providing a solid data foundation for subsequent intelligent inference and strategy generation. Using JSON format for storage enhances the system's scalability and data exchange efficiency, facilitating integration with multi-source platforms.

[0056] For ease of understanding, examples, but not limited to: Suppose the HAZOP report is an isopropylbenzene HAZOP analysis record sheet, which records the relevant analysis content of process node 02 benzene supply and purification. The risk factor data in the report can be: The process node corresponds to benzene supply purification. Risk source types include media, equipment, and location. The consequence chain is used to describe the consequences of a bowtie-type accident, including the initial accident, derivative accidents, and secondary hazards. Medium-related risk sources include benzene. Medium-related risk sources are configured with corresponding physicochemical parameters, such as flash point, ignition point, boiling point or explosion limits. Equipment-related risk sources include secondary benzene processors. Location-related risk sources include potential leakage points of equipment, such as static sealing points of dry benzene circulation pumps, mechanical seals such as valves, pumps, welds, flanges, etc. Trigger prompts are used to describe information about the failure of protective measures, such as abnormal or failed inlet / outlet pressure differential indication, abnormal or failed remote flow indication of dry benzene circulation pump outlet flow, safety valve failure, and leakage monitoring alarm. Triggering conditions are used to describe the causes of accident deviations, such as: high outlet pressure of dry benzene circulating pump, weld or flange leakage, valve or pump leakage, back-end blockage, and personnel misoperation. The consequence chain is used to describe the consequences of a bowtie-type accident, including the initial accident, derivative accidents, secondary hazards, etc. For example, consequence chain 1: leakage → fire → explosion → personnel poisoning → environmental pollution, consequence chain 2: leakage → personnel poisoning → environmental pollution. In the example, these two consequence chains overlap, and fire and explosion can also lead to personnel poisoning and environmental pollution.

[0057] The risk factor data may also include safety barriers, which are used to describe existing protective measures, recommended measures, such as monitoring measures, control measures, or emergency measures.

[0058] In risk factor data analysis, data samples with minor consequences and low probability of occurrence can be filtered out.

[0059] In the above embodiments, the system classifies and maps risk information of different natures into entity nodes in a graph based on the accident evolution logic and structured semantic information, specifically including the following categories: Risk source node: Represents the starting material or equipment unit that constitutes the risk, such as flammable media, pressurized equipment, flange connection points, etc. Node attributes include flash point, explosion limits, density, design pressure, service life, etc. Triggering event node: Represents an abnormal operation or physical failure that leads to risk activation. Node attributes include event type, trigger description, and probability of occurrence. Accident consequence nodes: These represent the main manifestations and derivative trends after the accident is activated. Node attributes include thermal radiation intensity, leakage rate, shock wave overpressure, rupture time delay, etc. Accident impact points: Record target objects that may be damaged or exposed within the scope of the accident, such as surrounding buildings, personnel, and environmental blocks; Emergency resource nodes: These represent the human resources, equipment, or technical means required for an accident response, including sensors, fire brigades, medical resources, etc. Node attributes include response time, capacity, and scope of action.

[0060] By modeling each stage of an accident's evolution as a node, this application constructs an accident risk evolution graph with a clear logical structure, supporting dynamic deduction and response simulation. The node definition method described above accurately expresses each component of a risk event and its static attributes, effectively supporting accident scenario modeling and path analysis. Standardized node classification and attribute parameter assignment help improve the graph's data integrity and model adaptability, supporting flexible multi-scenario expansion and enhancing the scientific rigor and operability of accident prediction and response strategy design.

[0061] In the above embodiments, the triggering relationship indicates that the existence or state change of a node directly activates subsequent nodes, such as equipment failure triggering a leakage event; the derivation relationship indicates the cascading evolution process of accident consequences, with clear time progression characteristics, such as leakage causing fire, and fire causing explosion; the dependency relationship expresses the dependence of response or inhibition mechanisms on resources or preceding nodes, such as fire extinguishing operations depending on the allocation of fire-fighting resources; the influence relationship describes the physical or functional impact of accident results on external entities, such as shock waves affecting surrounding residential areas or areas with people.

[0062] In the above embodiments, to enhance the dynamic expressiveness of the model, each relation edge can be configured with additional attributes, including: time delay, spatial distance, probability value, and intensity level. For example, time delay is t=8min, which represents the response delay of the derived relation; spatial distance is 10m away from the leakage source; probability value is 0.7; and intensity level is heat flux level, high pressure level, etc.

[0063] The aforementioned definition of relational edges and attribute configurations enable the accident graph to move beyond static relationships and dynamically display the spatiotemporal evolution characteristics and mechanisms of action of accidents, thereby enhancing the adaptability and predictive ability of the simulation model to complex scenarios. Fine-tuning the accident triggering mechanisms and response dependencies helps improve the accuracy of decision support and the simulation effect.

[0064] In some embodiments, the accident analysis step S12 further includes: The failure rate of the risk source corresponding to the process node is obtained as the failure rate of the base event, and the data on the operator error rate and the historical fluctuation data of the process parameters are obtained as the failure rate of the intermediate event. The failure rate data is obtained through the Offshore Reliability Data (OREDA) database, the reliability data provided by the equipment manufacturer, and the equipment failure records of the industry / enterprise. The operator error rate data is determined through the industry operation error statistical report and the enterprise safety training and assessment records. The historical fluctuation data of the process parameters includes excessively high outlet pressure, differential pressure indication of inlet and outlet materials, and remote transmission of the outlet flow of the dry benzene circulation pump. Specifically, the frequency of key parameters (such as pressure and flow) exceeding the normal range can be calculated from the DCS / SCADA historical database as the probability of the "intermediate event". The initial accident is constructed based on the Fault Tree Analysis (FTA) model. The intermediate and bottom events associated with the initial accident are entered. The failure rate of the intermediate events is obtained by performing an OR operation based on the failure rate of the bottom events. The annual occurrence probability of the initial accident is obtained by performing an OR operation based on the failure rate of the intermediate events. Identify the derivative accidents and / or secondary hazards associated with the initial accident, configure trigger probabilities for the derivative accidents and / or secondary hazards based on industry experience or expert judgment, and calculate the annual occurrence probability of the derivative accidents and / or secondary hazards based on the annual occurrence probability of the initial accident and the trigger probabilities.

[0065] Based on the above steps, a fault tree (FTA) model is introduced to quantitatively model various risk events in the accident evolution process. Risk sources associated with process nodes are used as base events, and failure rate data are obtained from multiple data sources, including the OREDA (Offshore Reliability Data) database, reliability manuals provided by equipment manufacturers, and industry or enterprise equipment operation and maintenance records. Subsequently, key influencing factors at the operational and process levels are extracted as intermediate events; the initial accident is used as the top event, and its corresponding intermediate and base events are identified. An OR gate is used to synthesize the base events to obtain the failure rate of the intermediate events; then, an OR gate is used to synthesize the intermediate events again to obtain the annual probability of the top event (i.e., the initial accident). Accident risk analysis shifts from qualitative speculation to quantitative calculation, improving the scientific rigor and reliability of risk modeling. The FTA model allows for systematic analysis of the accident causal chain, clarifying the failure contribution of equipment, personnel, and process parameters at each level, facilitating the precise formulation of control strategies. The wide range and traceability of data sources enhance the model's practical operability. Furthermore, by calculating the annual probability of occurrence in a hierarchical manner, the simulation model is equipped with quantitative support, thereby improving the credibility and prediction accuracy of emergency drills.

[0066] To facilitate understanding of how the annual probability of the top event is calculated based on the fault tree computation model, the embodiments of this application are illustrated below with examples.

[0067] Taking the initial accident as a benzene leak as an example, the intermediate and final events are broken down based on the triggering conditions and consequence chain, as follows: Intermediate Event A: High outlet pressure of dry benzene circulation pump, annual failure probability 12 times / year, annual fluctuation rate 0.0041 times / record; Intermediate event B: Back-end blockage (valve internal leakage closure → bottom event B1: outlet valve malfunction closure, failure rate 0.58 times / year; pipeline blockage → bottom event B2: pipeline impurity deposition, failure rate 0.05 times / year) → OR gate logic: If either B1 or B2 occurs, it will trigger back-end blockage (B = B1 ∨ B2), annual failure rate P(B) = P(B1) + P(B2) - P(B1) × P(B2) == 0.601 times / year Intermediate event C: Sealing point failure (leakage occurs) Bottom event C1: Leakage at static sealing points (welds / flanges) (e.g., failure rate 0.33 + 0.05 = 0.38 times / year) Underlying event C2: Leakage at mechanical seal points (valves / pumps) (e.g., failure rate 0.58 + 0.40 = 0.98 times / year) Intermediate event C = C1 ∨ C2 (OR gate logic) Intermediate event D: Human error (e.g., accidentally closing the outlet valve → failure rate 0.18 times / year; accidentally adjusting the feed flow rate → failure rate 0.15 times / year) → OR gate logic (D=D1∨D2) Top event T = A∨B∨C∨D In the above embodiments, this application configures the trigger probability based on historical accident data, for example as follows: the derivative accidents caused by the initial accident benzene leakage include: fire, explosion, poisoning, and secondary hazards include environmental pollution; The probability of a fire triggered by benzene leakage encountering an ignition source (such as electrical sparks or open flames in the factory area) is taken as 0.3. The probability of a benzene vapor cloud reaching its explosive limit (1.2%~8.0%) and encountering an ignition source to trigger an explosion is taken as 0.15. The probability of human poisoning after a fire or explosion is 0.8, and the probability of environmental pollution after a benzene leak is 0.6.

[0068] In another embodiment, the present application may also configure correction coefficients for the trigger probability, and correct the trigger probability based on the process characteristics of the process node. For example, but not limited to, if benzene has a boiling point of 80.1°C, it is easy to volatilize at room temperature, and the vapor cloud forms quickly, which will accelerate the explosion. Therefore, the explosion correction coefficient is taken as a value greater than 1, such as 1.2.

[0069] In some embodiments, the accident analysis step S12 further includes: The occurrence delay and impact range of the initial and derivative accidents are predicted based on the consequence quantification analysis model, such as "lethal radius XX meters" and "serious injury radius YY meters". The consequence quantification analysis model includes at least the following: national standard calculation model, Gaussian diffusion model / heavy gas diffusion model SLAB, pool fire thermal radiation model, vapor cloud explosion thermal radiation model, jet fire calculation model, explosion model, etc.

[0070] The national standard calculation model is used to implement the method for determining the external safety protection distance of hazardous chemical production and storage facilities in GB / T 37243-2019. The input parameters of the heavy gas diffusion model SLAB are leakage medium, leakage rate, wind speed, temperature, atmospheric stability, leakage height, leakage container material, leakage hole size, and occupational exposure limit. The output parameters are the diffusion range over time, concentration, and the range that can lead to poisoning and death. This model can be used to analyze the occurrence delay and impact range of benzene leakage diffusion accidents. The input parameters of the pool fire thermal radiation model are the area of ​​the liquid pool formed by the leaked benzene, the thickness of the liquid pool, and the heat of combustion of the leaked substance. The output parameters are the fire thermal radiation flux and fire spread rate over time and different ranges. This model can be used to analyze fire accidents caused by leakage. The area of ​​the liquid pool can be obtained based on the product of leakage rate and leakage time. The explosion model adopts the TNO multi-energy method model or the BLEVE model. The input parameters include vapor cloud volume and fuel mass. The output results are the overpressure peak value and its radius of action. The explosion model is used to analyze explosion accidents caused by fire.

[0071] Using the above model, this application can combine different models according to the accident scenario type to predict the damage diffusion process of the initial accident and its derivative events (such as leakage, fire, explosion) in the spatial and temporal dimensions, forming dynamic data such as "thermal radiation intensity field at t=5min of the accident" and "lethal zone range at t=10min of the accident", which constitute an important input basis for the wargaming simulation model.

[0072] This implementation method introduces multiple engineering consequence analysis models to achieve scientific prediction of accident consequences, especially dynamic modeling of the impact range and response delay, thereby improving the accuracy and reliability of accident simulation. Different models are adapted to different types of accidents, providing technical support for multi-scenario and multi-source risk modeling; the input parameters are readily available and the output results have engineering interpretability, facilitating the formation of spatialized and quantitative evolutionary scenarios by combining with graphical models, thus improving the data support capability and scenario reproduction of wargaming simulations.

[0073] In practical applications, in addition to the standard models mentioned above, commercial software (such as PHAST, ALOHA, and FLACS) can be used to calculate the consequences of accidents. For some new media or special environmental scenarios, CFD simulation software can be used to construct three-dimensional dynamic models to improve spatial granularity accuracy.

[0074] In some embodiments, the simulation step S3 includes: Step S31 of the wargaming piece definition involves configuring emergency drill entities, accident scenarios, and accident sites, and constructing and defining wargaming pieces and their attributes based on the structured risk dataset. The wargaming pieces include physical pieces, accident pieces, and board pieces. Piece attributes include type, position, state, radius of influence, operating rules, and controlled area. During the definition process, fields from the structured risk dataset (such as accident ID, trigger time, impact range, and quantification parameters) are mapped to the piece objects to achieve information fusion modeling. In step S32 of the simulation rule construction, the accident type, environmental variables, and emergency resource configuration are input into the rule parameter configurator to generate a benchmark simulation rule library. The benchmark simulation rule library includes entity attribute evolution rules, accident evolution rules, and adversarial rules. Entity attribute evolution rules are used to describe the changes in the ability attributes of entity pieces, such as the changes in ability attributes with time, chessboard pieces (terrain, environment, resources), and the influence of accident pieces. Accident evolution rules are used to define how the accident situation evolves over time (such as the increase in leakage intensity and the spread of fire) and the changes in pieces, such as the leakage piece becoming the fire piece, based on the accident map and consequence analysis model. Adversarial rules define the adversarial rules between chess pieces, such as the adversarial influence rules between the rescue of medical rescue pieces and fire pieces or explosion pieces on entity pieces, and the number of rounds required for fire units to extinguish fire points. In the confrontation exercise step S33, based on the benchmark simulation rule base, the accident evolution and emergency confrontation process are simulated through a round-based, real-time event-driven hybrid simulation algorithm and a random event generator, and the wargame simulation time series dataset is output.

[0075] In one embodiment, the simulation algorithm uses N minutes as the simulation step size, and each simulation round is processed in the following order: automatically advance and update the status of the accident pieces according to the simulation rules; receive action instructions from the physical pieces and execute the effects; determine whether the triggering conditions are met to generate new accident pieces; apply a random event generator to introduce sudden events (such as new faults, missing persons, sudden changes in wind direction, etc.) to affect the current confrontation strategy; record the status of all pieces, response actions, and accident situation in the current round, and write them into the wargame simulation time series dataset.

[0076] The simulation can continue until all incident situations stabilize or resources are exhausted, at which point the system will finally output complete data on the confrontation exercise process, including evolution paths, response behaviors, and effectiveness evaluation indicators.

[0077] By employing wargaming modeling and turn-based drills, this implementation transforms traditional static emergency drills into an interactive and evaluable dynamic simulation system, enhancing the understanding and response capabilities to the evolution of complex accidents. The introduction of a dual definition mechanism for pieces and rules effectively organizes various resources and scenario elements, improving modeling flexibility. The integration of a random event mechanism enhances the handling of uncertainties in the simulation, making the drills more closely resemble real-world scenarios. Simultaneously, the system can output standardized time-series datasets, facilitating subsequent evaluation and post-mortem analysis, and providing data support for emergency management decision-making.

[0078] The entity pieces include pieces representing corresponding emergency drill entities, which include rescue and response personnel, vehicles, and equipment. Each entity piece is configured with basic attributes, location attributes, capability parameters, and exclusive permissions. Basic attributes include role type (e.g., firefighter, medical team, rescue personnel) and survival value. Location attributes include coordinates and movement speed. The entity piece's capability parameters include firefighting ability, casualty rescue ability, and material transport ability. Coordinates include initial position and current coordinates. Firefighting ability includes firefighting efficiency value, effective range, and duration. Casualty rescue ability includes rescue capacity and rescue radius. Material transport ability includes carrying capacity and support range.

[0079] For ease of understanding, the following examples of physical chess pieces are provided in this application: Piece: "Water Tanker Fire Truck", Attributes: Coordinates (X,Y), Speed ​​3m / s, Fire Extinguishing Efficiency 40L / s, Range 65m, Continuous Spraying for 20 Minutes; Piece: "Medical Team", Attributes: Rescue capacity 2 people / time, response time 5 minutes, coverage radius 100m.

[0080] The accident pieces are constructed based on the structured risk dataset to represent accident nodes and evolution states, such as leak sources, ignition points, and explosion centers. Each accident piece is configured with a state switching attribute, such as switching from leakage to combustion, and from combustion to explosion. Optionally, the accident pieces' attributes also include: accident piece name / accident ID, accident center coordinates, type, state (triggered - not triggered), triggering condition, triggering probability, diffusion rate, derivative consequences (converted to other accident pieces), hazard level, and quantified attribute parameters (such as leakage rate, explosion overpressure, and toxicity concentration). A passage speed penalty coefficient is determined based on the influence range. This coefficient indicates that when a passage path passes through the influence range, the passage speed is inversely proportional to the distance between the passage point and the accident center. The accident piece subsequently appears or disappears in a consequence chain.

[0081] For ease of understanding, the following example of an accident chess piece is provided in the embodiments of this application: Piece: "Benzene Leakage", Attributes: ID=AX001, Coordinates (120,80), Leakage Rate 0.3m³ / h, State = Triggered, Probability = 5.7%, Evolution Chain: Benzene Leakage → Benzene Fire → Benzene Explosion, Hazard Level = Medium, Diffusion Rate is dynamically calculated by the SLAB model; Impact range setting: Define speed penalty or survival value penalty as a function of distance. If the path crosses the accident impact area, the passage speed will decrease inversely with the distance from the center point.

[0082] The chessboard and chess pieces include controlled areas, safe areas, and resource areas of the training ground. Each chessboard and chess piece is configured with attributes. Controlled areas are areas where passage is prohibited or damage is caused, such as fire zones. Their attributes include: coordinate range, passage penalty value, damage value, and time window. Safe areas allow for evacuation or centralized evacuation of personnel. Their attributes include: coordinate range, protection level, open duration, and maximum capacity. Resource areas provide emergency resources or capability enhancements, such as emergency supply depots, fire hydrants, fire monitors, and fire dikes. Their attributes include: resource type (supplies, water), enhancement value, and target.

[0083] For ease of understanding, the chessboard and chess pieces are illustrated in the following embodiments of this application: Piece: "Emergency Supplies Area", coordinates (240, 310), material type = fire monitor, effective range = 55m, fire extinguishing capacity bonus +10%. The effective range here is determined by centering on the accident point and using the fire monitor's protection radius as the effective range, thus defining a circle as the effective range of the fire monitor. Chess piece: "Emergency Shelter", coordinates (100, 200), capacity 50 people, protection level II, effective time 30 minutes.

[0084] Based on the above configuration, this application embodiment constructs a mapping table between HAZOP risk factors and wargame piece attributes through multi-dimensional attribute mapping, realizing dynamic attribute configuration: for example, the leakage rate is set to 0.5~2m³ / h and can be adjusted through a slider interface; the capabilities of physical pieces and the impact range of accident pieces are interactively calculated to form an actual operable space; collision judgment, influence effect and status linkage are supported between different pieces.

[0085] This implementation method achieves high-fidelity transformation of emergency drill elements from data to models, enabling wargaming models to possess refined and object-oriented modeling capabilities. By constructing entity, accident, and spatial environment pieces in a layered manner, it realizes dynamic response, spatial coordination, and resource scheduling during the drill process, significantly enhancing the interactivity and scenario realism of the drill. Simultaneously, multi-dimensional parameters support drill strategy simulation, effectiveness evaluation, and scheme optimization.

[0086] In some embodiments, the preset rule parameter configurator of this application automatically generates a deduction rule library with logical conditions and action results based on input parameters such as accident type, exercise environment and resource configuration, for the deduction engine to call.

[0087] The following are examples illustrating some typical deduction rules: For example, the evolution and countermeasures of benzene leak accidents: Triggering conditions and diffusion risk control: When the "benzene leak chess piece" is activated, if a "physical chess piece - personnel chess piece" enters the leak area, the system determines whether it carries an "anti-static tool chess piece". If not, the probability of generating an "static spark chess piece" increases by 5% each round, and this spark chess piece may trigger the activation of the "benzene fire chess piece". All "personnel chess pieces" located within the leak area must undergo a poisoning exposure assessment each round.

[0088] Accident termination condition determination: If all of the following conditions are met, the "benzene leak detection alarm" status will switch to "stop evolution": the "leak detection alarm" status changes from "high alarm" to "low alarm" or "not triggered"; the "operator" successfully closes the feed valve; and the "rescue personnel" completes the sealing operation.

[0089] For example, the rules for coordinated benzene fire suppression and protection include: Response dispatch and arrival time calculation: After the "Benzene Fire Chess" is activated, the nearest fire brigade is dispatched and it is calculated whether it can reach the accident point within 5 minutes. The arrival time of other fire units is linearly related to the distance.

[0090] Dynamic rules for fire extinguishing effect: After the "Fire Brigade Chess" and "Firefighter Chess" arrive, they reduce the heat radiation intensity of the "Benzene Fire Chess" by 20% each round. After 5 rounds, if the cumulative reduction reaches 0, the "Benzene Fire Chess" status changes to "Extinguished".

[0091] Terrain protection bonus rule: If the "fire dike" or "fire cannon" is in the "active" state, the rate of heat radiation growth in its effective area will be reduced by 50%, thus enhancing the fire extinguishing effect.

[0092] Personnel evacuation rules: "Evacuation operation workers" must evacuate to the "safe zone" within N rounds, otherwise their survival value will be deducted each round due to exposure to the heat radiation area.

[0093] Accident control determination: If the total value of the fire extinguishing capability parameters of the "fire and rescue team" is greater than or equal to the total value of the current fire heat radiation parameters, and the leak point is sealed, then the "benzene fire chess" status will switch to "stop evolution".

[0094] For example: Rules for responding to poisoning incidents: Using a toxic gas exposure assessment and intervention strategy, when the "fire and explosion chess" is activated, if the area concentration is >32ppm, the risk of poisoning for the "personnel chess" increases by 15% per round, and the "medical rescue chess" needs to be dispatched for intervention.

[0095] Medical Piece Cooperative Treatment Rules: After a "Medical Rescue Piece" reaches the poisoned area, it can reduce the poisoning risk of surrounding "Personnel Pieces" by 30% each round. It needs to continue for 2 rounds to complete one effective treatment.

[0096] Rules for using the "Poison Gas Adsorption Chess" in conjunction with the "Medical Rescue Chess": If the "Poison Gas Adsorption Chess" and the "Medical Rescue Chess" work together, the effectiveness of treatment will be increased to 90%.

[0097] Survival value change mechanism: Within the influence range of "Poisoned Chess", mild poisoning (30-50ppm): survival value -10 per round; moderate poisoning (50-300ppm): survival value -30 per round; severe poisoning (>300ppm): survival value -60 per round; those receiving medical treatment: survival value +25 per round (recovery mechanism), with 10 minutes of exposure within the corresponding influence range constituting one round.

[0098] For example: resource competition mechanisms and site allocation rules, including: Safe Zone Capacity Management Rules: The "Safe Zone Chess" accepts "Personnel Chess" based on the set time window and zone capacity. If the capacity is exceeded, subsequent personnel must wait for the next round before being considered for entry.

[0099] Resource Zone Priority Use Rule: Each "Resource Zone Piece" can only be used by the earliest arriving "Entity Piece" per round, and will gain corresponding ability bonuses, such as fire extinguishing efficiency +10% or medical capacity +1. Other pieces must wait until the next round to compete for resources.

[0100] In another embodiment, the rule parameter configurator can parse the input dataset and configuration items, and extract key variables, such as... Accident characteristic variables: accident type, triggering conditions, evolution path, evolution probability, scope of impact, etc.; Environmental variables: meteorological information (wind speed, wind direction, temperature), spatial structure (regional boundaries, distribution of resource points); Resource configuration parameters: rescue team type, number, initial coordinates, response capability, speed, etc.; Scene status indicators: current status of various chess pieces, available actions, skill trigger conditions, etc.

[0101] The preset rule parameter configurator of this application has built-in sets of multiple rule templates. Based on the extracted variables, the parameter configurator calls the corresponding template, fills in the parameter fields, and generates rule statements, forming a structured rule object. The rules include: Condition (if) → Triggering mechanism, prerequisite state, environment matching; Actions (then) include state transitions, ability changes, increases or decreases in survivability, and termination of incidents. Within the time frame → trigger duration, duration of effect (rounds), cooldown period, etc.

[0102] Then, the rule parameter configurator performs cause-effect graph modeling on the rules, connecting the triggering events, condition judgments and result states in a graph structure to form a state transition graph.

[0103] For example: "Firefighter arrives" + "Firefighting capacity value ≥ demand" → "Benzene fire status = Extinguished"; Alternatively: "Operator + Alarm + Successful Sealing" → "Benzene Leakage Stops Evolution"; Therefore, it can automatically detect logical loops, prevent infinite loops or triggering conflicts, and support the setting of parallel rules for multi-path triggering.

[0104] Each type of template corresponds to different training scenario requirements, and the rule template set is as follows: Accident evolution template: describes the dynamic process from the initial accident to the derivative / secondary accident; Adversarial response template: Describes the effects and conditions of resource-side entity pieces intervening in the incident piece; Site restriction templates: such as safe zone opening mechanisms, resource zone capacity restrictions, and the effects of toxic gas propagation and adsorption.

[0105] All the above generated rules are structured and output as standardized rule objects, including: rule ID, target object (chess piece number), triggering condition, execution action, scope of application, and effective time; these can be used for inference and execution in the event-driven module of the wargaming engine.

[0106] The rule parameter configurator uses structured input parsing, template matching, condition-action modeling, and random mechanism injection to form a complete set of inference rules covering the entire process of accident development and resource response behavior. This rule base provides the wargaming system with the basis for action and a state control framework, realizing a logical closed loop from structured risk data to dynamic exercise behavior, ensuring that the exercise simulation has a clear causal chain, complete decision-making conditions, and controllable emergency paths.

[0107] In some embodiments, the random event generator uses an improved Monte Carlo algorithm to introduce the probability distribution of historical drill accidents to generate sudden interference events. The specific process may include: Establish a disturbance event database, including but not limited to the following types: secondary equipment failure (such as fire pump failure, monitoring communication terminal or damage), personnel status changes (such as key personnel fatigue, loss of contact, misoperation), sudden changes in environmental factors (such as sudden changes in wind speed, air pressure fluctuations), changes in resource availability (such as channel blockage, communication interruption), and external uncertain events (such as lightning strikes, earthquakes, and linked off-site explosions). For each type of disturbance event, a trigger probability distribution model is established, and the probability density function can be set using historical accident data or expert experience. For example: the normal distribution is used to simulate the fluctuation trend of operational errors over time; the Poisson distribution is used for sparse but sudden equipment failures; the exponential distribution simulates sudden and serious accidents such as high-pressure vessel failures; and the uniform distribution is used for general disturbances in environmental variables. In each round of simulation, one or more random sampling operations are performed on each type of event: a time window and event evaluation frequency are set, and a random number generator is used to extract sample values ​​from each distribution model. If the sampling result falls within the preset trigger threshold range, the corresponding disturbance event is instantiated as a chess piece / event node and injected into the current simulation process. Once a random event is triggered, it will interfere with the existing wargame piece attributes, simulation status, or rule parameters. For example: the state of the accident piece changes abruptly (the leakage area expands, the heat flux increases); the physical piece fails to respond, its position is restricted, or its capabilities decrease; a new accident source is generated, changing the simulation path.

[0108] Generating random perturbation events through Monte Carlo simulation not only enhances the dynamic complexity and realism of wargames but also systematically verifies the stability and response effectiveness of contingency strategies under changing conditions. This mechanism combines flexibility and controllability, allowing for adjustments to the intensity and frequency of perturbations based on the exercise objectives, while also providing quantitative data support for subsequent risk assessment and algorithm optimization.

[0109] In some embodiments, the hybrid inference algorithm simulates the spatial diffusion of accident events (such as leaks, toxic gases, and fires) based on difference equations, and its calculation model is shown below: , For the first The radius of impact of the accident in each round. The diffusion coefficient of the leaking or diffusing medium. This is a cumulative time, in minutes. The wind speed is used as the reference. The increment of the influence radius is calculated each round, and the influence range of the accident piece is dynamically adjusted in this way. The influence range is also matched with the position of the actual piece in real time to determine whether the piece is in the exposed area.

[0110] In some embodiments, the hybrid inference algorithm establishes a conversion relationship between the wargaming system and the real geographic space. The chessboard is set with the accident point as the origin, the chessboard grid spacing = 10m, the horizontal axis is the x-axis (longitude conversion), and the vertical axis is the y-axis (latitude conversion). All pieces are bound to coordinates (x,y). For example, the coordinates of the "fire rescue team" in the 3rd round are (3,2), which corresponds to the real location (30m, 20m). The coordinates of the physical pieces are updated every round according to the movement speed. The position of the accident piece is fixed, and the spread range expands dynamically from the origin.

[0111] The hybrid inference algorithm executes the following process in each round: Update the status and impact range of the accident chess piece (based on the diffusion model); Execute entity chess actions (such as moving, extinguishing fires, and rescuing) according to the rule base; Call the random event generator to determine if a sudden disturbance has occurred; Determine whether the accident status has changed (e.g., cessation of evolution, state transition); Write the timing data and advance to the next round. That is, after each round of simulation, output a standardized timing data.

[0112] In some embodiments, each time series data point in the wargaming simulation time series dataset includes: The event includes a timestamp, event type, scope of impact, list of affected pieces, and list of unaffected pieces. The timestamp includes the number of rounds and the cumulative time.

[0113] This implementation combines turn-based and spatial diffusion calculations to achieve dynamic simulation of accident situations and refined tracking of entity response behaviors, enhancing the realism, data-driven nature, and assessability of emergency drills. Through standardized temporal data structures and spatial coordinate systems, a two-way mapping bridge is established between the wargaming system and real-world scenarios, facilitating the reuse of drill scenarios, data mining, and strategy verification.

[0114] The emergency drill simulation method using wargaming provided in this application combines several key technologies, including structured risk data modeling, graphical accident evolution representation, spatial scene mapping, rule-driven adversarial simulation, and standardized data output. This effectively solves the technical bottlenecks in traditional emergency drills, such as static scene settings, simplistic evolution processes, insufficient resource scheduling logic, and difficulty in quantifying and evaluating results. It has the following significant technical effects: By introducing text semantic parsing and accident graph modeling mechanisms, key elements in risk reports such as HAZOP are automatically extracted to generate structured risk datasets. Dynamic scenario models containing accident sources, triggering conditions, consequence chains, and affected areas are constructed, significantly improving the realism, professionalism, and efficiency of emergency drill scenario construction.

[0115] Based on diffusion models and difference equations, the impact range of accident scenarios (such as leaks, fires, and toxic gases) is dynamically calculated. Combined with a round-based progression mechanism, the evolution of accidents in the time-space dimension is accurately depicted. It supports multiple rounds of state switching and condition triggering, effectively improving the dynamic expressiveness and complexity simulation capabilities of accident drills.

[0116] Construct various entity chess piece models (such as fire fighting teams, medical teams, and rescue teams) and their response rules, and set resource scheduling conditions and termination judgment mechanisms in conjunction with accident triggering logic. Support complex interactions such as multi-role and multi-path collaborative response, resource seizure, delayed response, and capability confrontation to enhance the adversarial and strategic nature of the exercise.

[0117] By adopting a two-way mapping method between chessboard coordinates and actual spatial coordinates, the exercise pieces are precisely bound to physical spaces such as accident sites, resource areas, and safety zones. This supports the visualization and reproduction of exercise results in the map environment and can be extended to a GIS platform to achieve linkage between the simulation system and the actual scenario deployment.

[0118] Each round generates standardized time-series data, recording accident evolution parameters, entity status changes, capability usage trajectories, etc., forming CSV format data files and animations to display the results. This supports the reuse needs of subsequent exercise evaluation, strategy comparison, system training, etc., and provides quantitative support for emergency management.

[0119] A random event generation mechanism based on Monte Carlo simulation is introduced to dynamically trigger disturbance events (such as secondary leakage, personnel misoperation, environmental changes, etc.) according to the accident type and environmental variables, thereby enhancing the system's response and fault tolerance to uncertain risks and improving the realism and resilience verification capabilities of the exercise.

[0120] In summary, the simulation and deduction method provided in this application systematically enhances the technical capabilities of emergency drills in multiple dimensions such as data modeling, dynamic evolution, resource coordination, spatial adaptation, and result quantification, providing a complete and feasible technical solution for realizing a digital, intelligent, and combat-ready emergency drill system.

[0121] It should be noted that the steps shown in the above process or in the flowchart of the accompanying figures can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0122] This embodiment also provides an emergency drill simulation system. This device is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the terms "module," "unit," "subunit," etc., can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0123] Figure 5 This is a structural block diagram of an emergency drill simulation system according to an embodiment of this application, such as... Figure 5 As shown, the device includes: The report parsing unit is used to obtain the HAZOP report and parse the risk element data in the HAZOP report; The dataset acquisition unit is used to construct a risk evolution path map based on the risk element data, store the risk evolution path map as structured data, and obtain a structured risk dataset. The structured risk dataset is used to clarify the relationship between accident sources, derivative paths, and influencing factors. The structured risk dataset includes accident ID field, triggering condition field, triggering time field, derivative consequences field, evolution probability field, impact range field, and quantitative attribute parameter field. The simulation unit is used to map chess pieces based on the structured risk dataset, construct a wargame simulation model based on preset evolution rules, generate a preliminary scenario and perform dynamic simulation, and output a wargame simulation time series dataset.

[0124] It should be noted that the above modules can be functional modules or program modules, and can be implemented through software or hardware. For modules implemented through hardware, the above modules can reside in the same processor; or the above modules can be located in different processors in any combination.

[0125] In addition, combined Figure 1 The emergency drill simulation method using wargaming described in this application embodiment can be implemented by a computer device. The computer device may include a processor and a memory storing computer program instructions.

[0126] Specifically, the processor may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0127] The memory may include a large-capacity storage device for data or instructions. For example, and not limitingly, the memory may include a hard disk drive (HDD), a floppy disk drive, a solid-state drive (SSD), flash memory, an optical disk drive, a magneto-optical disk drive, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory may include removable or non-removable (or fixed) media. Where appropriate, the memory may be internal or external to the data processing device. In a particular embodiment, the memory is non-volatile memory. In a particular embodiment, the memory includes read-only memory (ROM) and random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), an electrically alterable read-only memory (EAROM), or flash memory, or a combination of two or more of these. Where appropriate, the RAM can be Static Random-Access Memory (SRAM) or Dynamic Random-Access Memory (DRAM). DRAM can be Fast Page Mode Dynamic Random-Access Memory (FPMDRAM), Extended Data Out Dynamic Random-Access Memory (EDODRAM), Synchronous Dynamic Random-Access Memory (SDRAM), etc.

[0128] Memory can be used to store or cache various data files that need to be processed and / or communicated, as well as possible computer program instructions executed by the processor.

[0129] The processor reads and executes computer program instructions stored in memory to implement any of the emergency simulation methods for applying wargaming in the above embodiments.

[0130] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0131] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. An emergency drill simulation method using a war game, characterized by, Comprising: a report analysis step of obtaining a HAZOP report and analyzing risk element data in the HAZOP report; the report analysis step comprises: a text semantic analysis step of extracting risk element data in the HAZOP report based on a pre-trained text analysis model, the risk element data including process nodes, risk sources, trigger prompts, trigger conditions, consequence chains, and safety barriers, the consequence chain including initial accidents, derivative accidents, and secondary hazards; an accident analysis step of determining the occurrence probability of initial accidents, derivative accidents, and secondary hazards in the consequence chain based on the trigger conditions and the consequence chain in the risk element data, and predicting the occurrence time delay and impact range of accidents in the consequence chain based on a consequence quantification analysis model; a data set acquisition step of constructing a risk evolution path graph based on the risk element data, storing the risk evolution path graph as structured data, and obtaining a structured risk data set; the data set acquisition step further comprises: a graph definition step of creating entity nodes and relationship edges based on the risk element data, mapping to generate a risk evolution path graph, the entity nodes in the risk evolution path graph including risk sources, trigger events, accident consequences, accident impacts, and emergency resources, and the relationship edges including trigger relationships, derivative relationships, dependency relationships, and impact relationships, and configuring relationship attributes for entity relationships in combination with the occurrence time delay and the impact range; a data set generation step of converting the risk evolution path graph into structured data to obtain a structured risk data set, the structured risk data set including an accident ID field, a trigger condition field, a trigger time field, an accident consequence field, an evolution probability field, an impact object field, an impact range field, and a quantification attribute parameter field; a deduction simulation step of mapping chess pieces based on the structured risk data set and constructing a war game deduction model based on a pre-set evolution rule to perform deduction simulation and output war game deduction time series data; the deduction simulation step comprises: The steps for defining wargame pieces include configuring emergency drill entities, accident scenarios, and accident sites, and constructing and defining wargame pieces and their attributes based on the structured risk dataset. The wargame pieces include entity pieces, accident pieces, and board pieces. Entity pieces include pieces representing corresponding emergency drill entities, which include rescue and response personnel, vehicles, and equipment. Basic attributes, location attributes, capability parameters, and exclusive permissions are configured for each entity piece. Basic attributes include role type and survival value; location attributes include coordinates and movement speed; capability parameters include firefighting ability, casualty rescue ability, and material transport ability; coordinates include initial position and current coordinates; firefighting ability includes firefighting efficiency value, effective range, and duration; casualty rescue ability includes rescue capacity and rescue radius; and material transport ability includes carrying capacity. The accident pieces, constructed based on the structured risk dataset, represent accident nodes and evolution states, and are configured with state switching attributes. A passage speed penalty coefficient is determined based on the accident piece's influence range. This coefficient indicates that when a passage path passes through the influence range, the passage speed is inversely proportional to the distance between the passage point and the accident center point. The accident pieces appear or disappear along with the consequence chain. The chessboard pieces include controlled zones, safe zones, and resource zones in the training area, and are configured with attributes. Controlled zones prohibit passage or cause damage; their attributes include: coordinate range, passage penalty value, damage value, and time window. Safe zones allow for evacuation or centralized personnel transfer; their attributes include: coordinate range, protection level, opening duration, and maximum capacity. Resource zones provide emergency resources or capability enhancements. The simulation rule construction steps involve inputting accident type, environmental variables, and emergency resource configuration in the rule parameter configurator to generate a benchmark simulation rule library, which includes entity attribute evolution rules, accident evolution rules, and adversarial rules. The adversarial exercise steps, based on the aforementioned benchmark simulation rule base, simulate the evolution of accidents and emergency confrontation processes using a round-based, real-time event-driven hybrid simulation algorithm and a random event generator, outputting the wargame simulation time series dataset. The random event generator uses an improved Monte Carlo algorithm to introduce the probability distribution of historical exercise accidents to generate sudden interference events. The simulation algorithm has a simulation step size of N minutes, and each round of simulation is processed in the following order: automatically advance and update the state of the accident piece according to the simulation rules; receive and execute the action instructions of the physical piece; determine whether the triggering conditions are met to generate a new accident piece; apply the random event generator to introduce sudden interference events to affect the current confrontation strategy; record the state of all pieces, response actions, and accident situation in the current round, and write them into the wargame simulation time series dataset. The random event generator uses an improved Monte Carlo algorithm to introduce the probability distribution of historical simulation accidents to generate sudden interference events. This includes: establishing a disturbance event library, including the following types: secondary equipment failure, personnel status change, sudden environmental factor change, resource availability change, and external uncertain time; establishing a trigger probability distribution model for each type of disturbance event; and performing one or more random sampling operations for each type of event in each simulation round. The random number generator extracts sample values ​​from each distribution model. If the sampling result falls within the preset trigger threshold range, the corresponding disturbance event is instantiated as a piece / event node and injected into the current simulation process.

2. The method of claim 1, wherein the method is characterized by, The accident analysis steps further include: The failure rate of the risk source corresponding to the process node is obtained as the failure rate of the bottom event, and the data on the operator's operation error rate and the historical fluctuation data of process parameters are obtained as the failure rate of the intermediate event. The initial accident is constructed based on the fault tree calculation model. The intermediate and bottom events associated with the initial accident are entered. The failure rate of the intermediate events is obtained by performing an OR operation based on the failure rate of the bottom events. The annual occurrence probability of the initial accident is obtained by performing an OR operation based on the failure rate of the intermediate events. Identify the derivative accidents and / or secondary hazards associated with the initial accident, configure trigger probabilities for the derivative accidents and / or secondary hazards, and calculate the annual occurrence probability of the derivative accidents and / or secondary hazards based on the annual occurrence probability of the initial accident and the trigger probabilities.

3. The emergency exercise simulation method using war game according to claim 2, wherein The accident analysis steps further include: The occurrence delay and impact range of the initial and derivative accidents are predicted based on the consequence quantification analysis model.

4. The method of claim 1, wherein the method is characterized by, Each time series data point in the wargaming simulation time series dataset includes: Timestamp, incident type, scope of impact, list of affected pieces, list of unaffected pieces.

5. An emergency exercise simulation system using a war game, which implements the emergency exercise simulation method according to any one of claims 1 to 4, characterized by, include: The report parsing unit is used to obtain the HAZOP report and parse the risk element data in the HAZOP report; The dataset acquisition unit is used to construct a risk evolution path map based on the risk element data, store the risk evolution path map as structured data, and obtain a structured risk dataset. The simulation unit is used to map chess pieces based on the structured risk dataset, construct a wargame simulation model based on preset evolution rules, and output a wargame simulation time series dataset.

6. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement an emergency drill simulation method for applying wargaming as described in any one of claims 1 to 4.

Images