Park safety risk early warning and response plan generation method

Abstract

The invention discloses a park safety risk early warning and response plan generation method, and relates to the technical field of knowledge maps, and the method comprises the steps: collecting multi-source data, and carrying out the preprocessing; constructing a knowledge graph to generate a triple structure of nodes, edges and attributes, integrating the triple structure and the preprocessed multi-source data through path query to obtain a risk path, evaluating the type, grade and probability of the risk path by using a semantic reasoning rule base and a random forest model, and generating a risk evaluation report; and analyzing the interaction influence of the plurality of risk paths through a graph neural network model to obtain an interaction risk probability, outputting a priority adjustment suggestion according to the interaction risk probability, the preprocessed multi-source data and a priority rule, allocating rescue resources, and generating an optimization plan. According to the method, the knowledge graph is constructed, and the triple structure of the nodes, the edges and the attributes is generated, so that accurate modeling of complex risk factors in the park and the mutual relation of the complex risk factors is realized.

Description

Technical Field

[0001] The present invention relates to the field of knowledge graph technology, and in particular to a method for generating a park safety risk early warning and response plan. Background Art

[0002] With the integration of information and intelligent perception technologies, campus safety risk warning has gradually developed into a multi-level linkage architecture; in terms of data collection, the modern Internet of Things achieves real-time acquisition of multi-source data through distributed nodes, and effectively eliminates cross-source interference through data cleaning; in the field of knowledge modeling, it focuses on building a structured risk association network, integrating equipment topology relationships, safety specification documents and historical event records, and ultimately forming a risk knowledge base with semantic attributes; in the risk assessment link, integrated learning models are widely used in risk probability prediction, combined with path search algorithms to identify key risk paths; however, there are still limitations in the dynamic coupling analysis of risk paths. Most of them use static association models, which make it difficult to fully consider the dynamic interactions between equipment, environment and human factors, resulting in weak real-time perception of the evolution of complex risks.

[0003] Existing technologies still face challenges in park safety risk warning: analysis of complex interactions between risk paths usually relies on traditional statistical methods or simple logical association models; for example, by establishing regression models to evaluate the correlation between risk factors, or using machine learning algorithms such as decision trees for classification predictions; these methods can often only provide an understanding of linear relationships, but it is difficult to capture complex nonlinear interactions and deep implicit patterns. Summary of the Invention

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

[0005] Therefore, the present invention provides a park safety risk early warning and response plan generation method to solve the analysis problem of the complex interactive impact between risk paths.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In the first aspect, the present invention provides a method for generating a park safety risk warning and response plan, which includes collecting and preprocessing multi-source data, wherein the multi-source data includes environmental data, personnel activity data, and equipment status data; constructing a knowledge graph to generate a triple structure of nodes, edges, and attributes, integrating the triple structure and the preprocessed multi-source data through path query to obtain a risk path, using a semantic reasoning rule library and a random forest model to evaluate the type, level, and probability of the risk path, and generate a risk assessment report; analyzing the interactive impact of multiple risk paths through a graph neural network model to obtain an interactive risk probability, outputting priority adjustment suggestions based on the interactive risk probability, the preprocessed multi-source data, and the priority rules, allocating rescue resources, and generating an optimized plan; distributing the optimized plan to the terminal device through the emergency management platform, displaying the plan content using a visual interface, and further optimizing the plan based on the semantic reasoning rule library to generate an operation log.

[0007] As a preferred solution of the park safety risk warning and response plan generation method described in the present invention, the construction of a knowledge graph to generate a triple structure of nodes, edges and attributes refers to collecting risk information from historical accident data and safety guidelines, extracting nodes and edges from the risk information using natural language processing technology, determining the causal relationship weights of the edges and the node risk threshold as attributes through statistical analysis, importing the nodes, edges and attributes into a graph database, and verifying the accuracy of the node risk threshold and edge weight by comparing historical accident data to form a structured knowledge representation that can be stored and queried.

[0008] As a preferred solution of the park security risk early warning and response plan generation method of the present invention, wherein: the node is an entity type; The edges are relationships between entity types; The attributes are the node risk threshold and the edge causality weight.

[0009] As a preferred solution of the park safety risk early warning and response plan generation method of the present invention, wherein: the entity type includes risk type, environmental factors, equipment status, personnel density, resources and spatial location; The edges include causal relationships, spatial relationships, and state dependency relationships.

[0010] As a preferred solution of the park safety risk warning and response plan generation method described in the present invention, the method of obtaining the risk path by integrating the triple structure and multi-source data through path query refers to activating the entity types that meet the node risk threshold based on the nodes, edges, attributes and multi-source data of the knowledge graph, and querying the causal relationship path of the edge connection through the path query function of the graph database to identify the risk path.

[0011] As a preferred embodiment of the method for generating a park safety risk warning and response plan according to the present invention, the semantic reasoning rule base generates reasoning rules based on the causal relationship of historical accident data and the safety threshold of safety criteria, compares the values ​​of multi-source data with the reasoning rules, and determines the type and level of the risk path; The random forest model refers to processing the numerical values ​​of multi-source data and the attributes of the knowledge graph through multiple decision trees, counting the decision tree results and predicting the risk probability of the risk path through voting.

[0012] As a preferred solution of the park safety risk warning and response plan generation method described in the present invention, the method of analyzing the interactive impact of multiple risk paths through a graph neural network model to obtain the interactive risk probability refers to mapping the risk path into a graph structure, using node embedding and message passing mechanisms to quantify the interaction between multiple risk paths and generate the interactive risk probability.

[0013] As an optimal solution for the park safety risk warning and response plan generation method described in the present invention, the priority rule is based on historical accident data and sets a priority threshold. When the interaction risk probability exceeds the priority threshold, a high-risk interaction scenario is determined and the priority rule is triggered.

[0014] In a second aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements any step of the campus safety risk warning and response plan generation method as described in the first aspect of the present invention.

[0015] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the campus safety risk warning and response plan generation method as described in the first aspect of the present invention.

[0016] The beneficial effects of the present invention are as follows: by constructing a knowledge graph and generating a triple structure of nodes, edges, and attributes, the present invention achieves accurate modeling of complex risk factors and their interrelationships within the park. It not only integrates historical accident data with real-time monitoring information, but also accurately captures the causal relationship between different risk factors. It uses a semantic reasoning rule library combined with a random forest model to assess the risk type, level, and probability of these paths, improving the accuracy and reliability of risk predictions and generating detailed risk assessment reports. It not only overcomes the limitations of traditional single-dimensional analysis, but also makes emergency response measures more reasonable, and can dynamically adjust plans, optimize resource allocation, and improve the speed and efficiency of responding to emergencies, ultimately achieving the goal of reducing park safety risks and protecting the safety of people and property. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 A flowchart of the method for generating a park security risk warning and response plan.

[0019] Figure 2 Processing flow chart for the core method.

[0020] Figure 3 Build a flowchart for the knowledge graph.

[0021] Figure 4 This is a risk path analysis flow chart. DETAILED DESCRIPTION

[0022] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0023] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0025] Reference Figures 1 to 4 This embodiment provides a method for generating a campus security risk warning and response plan, including the following steps: S1. Collect multi-source data and pre-process them.

[0026] Furthermore, IoT devices are deployed in chemical warehouses in the industrial park (such as Area A). These devices include infrared temperature sensors, methane gas detectors, 4K cameras, and industrial sensors to collect multi-source data, including environmental data, personnel activity data, and equipment status data. Infrared temperature sensors collect temperature and humidity from environmental data every second, covering the tops and floors of the warehouse’s storage units; Methane gas detectors collect combustible gas concentrations from environmental data every second and are deployed at the vents and entrances of warehouse storage units; The 4K camera collects the number of people and their locations (through face detection and coordinate positioning) in human activity data and the valve switch status (through image recognition) in equipment status data every second. Industrial sensors collect equipment status data every second, including pump operating power and motor vibration frequency, and are installed on pumps and motors in the warehouse; All IoT devices package environmental data, device status data, and personnel activity data into multi-source data and transmit them to edge computing nodes through the MQTT protocol, and temporarily store them in the memory of the edge computing nodes in JSON format.

[0027] Filter multi-source data through edge computing nodes, including deduplication and removal of invalid data; Deduplication removes duplicate records by comparing timestamps and data types in multiple source data; Invalid data is eliminated based on preset judgment thresholds. For example, records with temperature exceeding -20°C to 100°C, humidity exceeding 0% to 100%, gas concentration exceeding 0% to 5%, pump power exceeding 0 to 50 kilowatts, vibration frequency exceeding 0 to 100 Hz, valve switch status not 0 or 1, number of people exceeding 0 to 100, and location exceeding the range of Warehouse Area A (0≤x≤100 meters, 0≤y≤50 meters) are considered invalid data and discarded directly. The preset judgment thresholds are based on the Regulations on the Safety Management of Hazardous Chemicals. The output is filtered multi-source data, which is stored in the local cache of the edge computing node.

[0028] Obtain filtered multi-source data from the local cache of the edge computing node and use the mean method to fill in missing values; The mean method checks the filtered multi-source data one by one to identify missing records where the numeric fields are empty or marked as invalid. For example, a temperature sensor records no data or a camera records an empty number of people. For each missing record (after filtering multi-source data), the mean method extracts normal records of the same type and location (with non-empty values ​​and not marked invalid) from the previous 5 minutes based on the multi-source data type (for example, temperature) and location (for example, Warehouse Area A). The mean is then calculated as the fill-in value. For example, if a temperature sensor is missing temperature data for a period of time, and the temperature sensor at the same location generates five normal records in the previous 5 minutes, which are 50°C, 51°C, 49°C, 50.5°C, and 50.2°C, the fill-in value is 50.14°C. If there are no normal records in the previous 5 minutes, the value field is left empty. The number of people and locations in the personnel activity data are not filled due to their discreteness. The missing records keep the value field empty. For example, if a camera records the number of people but it is empty, the value is kept empty. The output is the padded multi-source data, which is stored in the local cache of the edge computing node.

[0029] A sliding window method was used with a 5-second time window to calculate moving averages for temperature, humidity, and gas concentration in the padded environmental data, and pump power and vibration frequency in the padded equipment status data. For example, the gas concentrations were recorded at 0.11%, 0.12%, 0.13%, 0.10%, and 0.14%, with a moving average of 0.12%. The 5-second window was selected based on the per-second acquisition frequency of the padded multi-source data, the second-level noise characteristics of the environmental data and equipment status data, and the performance of the edge computing nodes. Experimental verification was conducted by testing 3-second, 5-second, and 10-second windows, comparing the smoothing effect (percentage reduction in standard deviation) and calculation latency, and confirmed that 5 seconds was the optimal value. The number of people and locations in the personnel activity data are not smoothed due to their discreteness, so the original values ​​are retained; The output is smoothed multi-source data in JSON format, including time, location, data type, value, and unit, and stored in the local cache of the edge computing node.

[0030] Obtain smoothed multi-source data from the local cache of edge computing nodes, unify the format, synchronize time (calibrate timestamps based on the NTP protocol), and group data (by location, such as warehouse area A, and data type (temperature, gas concentration, etc.)) to generate standardized multi-source data. The standardized multi-source data is uploaded to the cloud database (MongoDB) via the 5G network and stored as a JSON document containing time, location, data type, value, and unit.

[0031] S2. Construct a knowledge graph to generate a triple structure of nodes, edges, and attributes. Integrate the triple structure and preprocessed multi-source data through path query to obtain the risk path. Use the semantic reasoning rule library and random forest model to evaluate the type, level, and probability of the risk path and generate a risk assessment report.

[0032] Furthermore, we build a knowledge graph based on the graph database (Neo4j), as follows: Obtain historical accident data from the internal accident records of the chemical warehouse in the industrial park; Risk information is extracted from historical accident data and industry safety guidelines, including risks of fire (caused by high temperature and flammable gases), explosion (gas leakage), equipment failure (pump overload) and personnel safety (excessive density).

[0033] Extract nodes, edges, and attributes from risk information to construct the triple structure of the knowledge graph, as follows: Nodes represent entity types, including risk types (such as fire, explosion, and equipment failure), environmental factors (such as high temperature and excessive flammable gas concentration), equipment status (such as pump overload and abnormal vibration frequency), occupancy density (for example, the number of people exceeds the safety threshold, >30 people), resources (such as fire hydrants and emergency exits), and spatial locations (such as Warehouse Area A and Warehouse Area B). Environmental factors, equipment status, resources, and spatial locations are extended node content, rationally derived based on risk information triggering conditions, historical accident data, industry safety guidelines (such as the Regulations on the Safety Management of Hazardous Chemicals and the Fire Protection Guidelines for Building Design), and application scenario requirements. This extension is reasonable and is intended to support risk analysis and emergency response. Edges represent relationships between entity types, including causal relationships (e.g., high temperatures causing fires and excessive concentrations of combustible gases causing explosions), spatial relationships (e.g., warehouse area A is adjacent to warehouse area B), and state dependencies (e.g., pump failure accompanied by abnormal vibration frequency). Attributes include node risk thresholds (e.g., high-temperature fire risk threshold: temperature > 50°C, duration ≥ 5 minutes, set according to fire risk control guidelines for chemical storage environments in warehouse area A to prevent high-temperature fires) and edge causality weights (e.g., the probability of high temperature causing a fire is 0.8, based on historical accident data statistics); Natural language processing technology is used to parse text data of risk information and extract nodes and edges. For example, when parsing "high temperature caused fire" in historical accident reports, the nodes extracted are "high temperature" and "fire", and the edge is "triggered". Statistical analysis of causal relationship weights is used to extract attributes. For example, if 80 of 100 fire accidents were related to high temperature, the weight of high temperature causing fire is set to 0.8. The extracted nodes, edges, and attributes are imported into the graph database to generate the initial knowledge graph. For example, the node "high temperature" is connected to the node "fire" through the edge "triggered", and the edge attribute is "causal relationship with weight 0.8".

[0034] A verification process ensures that the node risk thresholds and edge weights of the initial knowledge graph are accurate. The verification process involves comparing the node risk thresholds (e.g., temperature > 50°C) and edge weights (e.g., 0.8 for high temperature fires) of the initial knowledge graph with historical accident data and calculating the deviation percentage. For example, if the edge weight of the initial knowledge graph is 0.8, and the historical accident data statistics are 0.78, then the deviation percentage is expressed as: (0.8-0.78) / 0.78×100%=2.56%; Among them, the node risk threshold and edge weight can only pass the verification if the deviation percentage is less than 5% (based on the industry error tolerance); After the verification process is completed, the output is the constructed knowledge graph, which is stored in the graph database.

[0035] Obtain the most recent standardized multi-source data from the cloud database and load the knowledge graph into the graph database; Retrieve nodes, edges, attributes, and risk thresholds from the graph database, extract JSON structured fields (time, location, data type, value, and unit) from standardized multi-source data, and match the JSON structured fields one by one to the entity types and risk thresholds in the knowledge graph, as follows: If the flammable gas concentration is ≥ 0.1% (based on the chemical safety storage guidelines), match the entity of the environmental factor: if the flammable gas concentration is too high, for example, if the gas concentration is 0.12%, then activate the entity of the environmental factor; If the pump power is ≥ 40 kW (80% of the rated power in the equipment manual), match the entity of the equipment status: Pump overload. For example, if the pump power is 42 kW, then activate the entity of the equipment status. If the number of people is ≥30 and the spatial location is Warehouse Area A (based on the personnel evacuation criteria and warehouse area), match the entity of personnel density: if the number of people is >30, for example, if the number of people is 35, then activate the entity of personnel density.

[0036] Use the path query function of the graph database to find the causal relationship between activated entities triggered by JSON structured field matching, forming a potential risk path. For example, "high temperature" is connected to "fire" through the edge "trigger", or "excessive combustible gas concentration" and "high temperature" jointly trigger "explosion", forming a risk path: high temperature + excessive combustible gas concentration → explosion.

[0037] The semantic reasoning rule base is used to evaluate the risk path, as follows: Extract causal relationships from historical accident data of risk information, such as high temperature leading to fire; obtain safety thresholds from chemical safety storage guidelines, such as temperature > 50°C and gas concentration ≥ 0.1%, which are set based on the chemical characteristics and historical accident data in warehouse area A; Generate inference rules based on causal relationships and safety thresholds and combine them with the nodes, edges, and attributes of the knowledge graph, and store them in the semantic inference rule library; The semantic reasoning rule library loads predefined reasoning rules; Extract JSON structured fields from standardized multi-source data, such as temperature 51°C and gas concentration 0.12%; Get a numerical value from a JSON field, such as 51°C, as the field value; Compare the field value to the conditions of the inference rule (e.g., temperature > 50°C); Confirm through comparison whether a risk path has been triggered, such as "high temperature → fire"; If the field value meets the conditions, the conclusion of the inference rule is output, such as the risk type is fire and the level is high; if it does not meet the conditions, the conclusion of the inference rule is not output.

[0038] The random forest model is trained based on historical accident data to predict the risk probability of the risk path, as follows: Extract JSON structured fields and field values ​​from standardized multi-source data, and obtain attributes from knowledge graphs; Normalize the field values ​​of the JSON structured fields and the attributes of the knowledge graph, map the field values ​​and attributes to the interval [0, 1], unify the dimensions to generate the risk feature vector, and represent the conditions of the risk path; Load a random forest model trained based on historical accident data. The training process is as follows: Based on the random forest algorithm, use the historical accident data to construct multiple decision trees (for example, 100 trees). Each tree obtains a subset from the historical accident data through random sampling with replacement. The subset includes features (temperature, gas concentration, and edge weights) and historical risk path labels (determined based on the event outcomes of the historical accident data, for example, "high temperature → fire" is 1 if it occurs and 0 if it does not). The subset is used to train the decision tree, generate node thresholds and branch paths, and form a decision tree structure to predict binary classification results. The risk feature vector is input into the random forest model, where each decision tree of the random forest model predicts whether the risk path will occur. By comparing the risk feature vector with the node threshold generated by training, the model outputs a binary classification result (1 for occurrence and 0 for non-occurrence, for example, "high temperature → fire" is predicted as 1); The binary classification results of each decision tree are counted, and the proportion of risks occurring is calculated through voting to generate the risk probability. The voting method is as follows: the number of decision trees that predict 1 among multiple decision trees is counted, and the number is divided by the total number of decision trees to obtain the risk probability. For example, if there are 100 decision trees in total, and 85 of them predict 1 for "high temperature → fire", then the risk probability is 85%.

[0039] Inference conclusions and risk probabilities are directly integrated into risk assessment data. The fields of inference conclusions, risk probability, and standardized multi-source data are merged into a JSON document containing time, location, data type, value, unit, risk type, risk level, risk probability, priority, impact scope, and risk path, and stored in the local cache of the cloud computing node. Obtain risk assessment data from the local cache of cloud computing nodes, aggregate high-risk records by risk level (high, medium, and low) and risk probability (e.g., <30% for low risk, 0%-60% for medium risk, and >60% for high risk), and generate aggregated results; Generate a risk assessment report based on the aggregated results; the risk assessment report includes risk type, risk level, risk probability, priority, impact scope, and risk path; The risk assessment report is uploaded to the cloud database via the 5G network and stored as a JSON document.

[0040] S3. Analyze the interactive impact of multiple risk paths through the graph neural network model to obtain the interactive risk probability. Based on the interactive risk probability, preprocessed multi-source data, and priority rules, output priority adjustment suggestions, allocate rescue resources, and generate an optimized plan.

[0041] Going a step further, risk assessment reports are loaded from the cloud database to filter out high-risk records; Generate risk plan templates based on historical accident data and store them in a graph database as an extension of the knowledge graph; The risk plan template includes a command structure (e.g., emergency command team), resource allocation (e.g., fire trucks and fire extinguishers), evacuation routes (e.g., emergency exit on the north side of warehouse area A), and communication procedures (e.g., notifying the fire department); Compare the fields of high-risk records with the risk plan template and match them with the appropriate risk plan template. For example, for a fire risk (e.g., a risk probability of 85%, high risk), match the risk plan template of "High-Level Fire Risk Plan" to generate initial response measures (e.g., activating fire trucks and evacuating personnel) and resource allocation (e.g., 2 fire trucks and 10 fire extinguishers); Resource entities and spatial location entities are obtained from the knowledge graph to support resource allocation and evacuation route planning in the risk plan template. For example, the fire truck is located at the "adjacent fire station (2 kilometers away from warehouse area A)" and the emergency exit is located at the "north side of warehouse area A".

[0042] Extract multiple risk paths from risk assessment reports; Using a pre-trained graph neural network model, we analyze the interactive impact of multiple risk paths and generate optimized plans, as follows: Multiple risk paths extracted from the risk assessment report are mapped into a graph structure using the knowledge graph's path query function to obtain the corresponding activated entities, adjacency relationships, and edge weights. The graph represents the risk paths, with activated entities as nodes, adjacency relationships as edges, and edge weights as edge attributes. Perform node embedding on the nodes in the graph representation, mapping each node to a low-dimensional vector to represent the entity's characteristics. For example, "high temperature" is mapped to a low-dimensional vector to represent the characteristics of "high temperature". Then, through the message passing mechanism, the low-dimensional vectors generated by node embedding are used to exchange information. Each node receives the low-dimensional vectors of neighboring nodes and performs weighted aggregation based on the edge weights. For example, the low-dimensional vector of "high temperature" is passed to "fire" through the edge "trigger" with an edge weight of 0.8. After receiving the low-dimensional vector of "fire", the low-dimensional vector of the node is updated. After multiple rounds of message transmission, node information is aggregated and mapped to interaction risk probabilities using the Sigmoid activation function. Interaction risk probabilities quantify the interactive impact of multiple risk paths, indicating the likelihood that two risk paths will occur simultaneously or exacerbate each other in the current scenario. For example, a 90% interaction risk probability for "fire" and "explosion" indicates a 90% probability of both occurring simultaneously or exacerbating each other in warehouse area A. Based on historical accident data, a priority threshold (e.g., 85%) is set. When the interaction risk probability exceeds the priority threshold, a high-risk interaction scenario is identified and priority rules are triggered. Priority rules include priority evacuation and priority deployment of rescue resources. Generate priority adjustment suggestions based on priority rules and interaction risk probabilities. For example, if the interaction risk probability of "fire + explosion" is 90%, which meets the priority rules, it is recommended to "prioritize evacuation of explosion risk areas"; Obtain multi-source data after real-time preprocessing, and based on optimization rules, multi-source data, interactive risk probability and priority adjustment suggestions, confirm the priority type (such as priority evacuation), allocate rescue resources, and generate an optimized plan.

[0043] Natural language generation technology is used to convert the optimized plan into clear instructions. The instructions include response measures (based on priority rule types, such as prioritizing the evacuation of explosion-risk areas), resource allocation (such as deploying fire trucks and explosion-proof equipment), priorities, map data (based on the spatial location entities and standardized multi-source data of the knowledge graph to generate regional map information) and resource location (based on the real-time multi-source data of resource entities in the knowledge graph to record the location information of rescue resources); and the instructions are uploaded to the cloud database via 5G.

[0044] S4. Distribute the optimized plan to the terminal device through the emergency management platform, use the visual interface to display the plan content, and further optimize the plan based on the semantic reasoning rule library to generate operation logs.

[0045] Furthermore, the optimization plan JSON document is extracted from the cloud database and converted into a response plan document, preserving the instruction, resource allocation, priority, map data, and resource location fields; Response plan documents are distributed via the 5G network to managers' mobile devices, fire department command equipment, and broadcast equipment in Warehouse Area A. The distribution process uses an encrypted communication protocol (AES-256) to ensure data transmission security. The manager's mobile device receives the response plan document, which displays instructions (such as "Prioritize evacuation of explosion risk areas") and resource allocation (such as three fire trucks and 10 emergency personnel), allowing the manager to confirm or make adjustments, such as adjusting resource allocation; The fire department's command equipment receives response plan documents, displays resource allocation and map data (such as warehouse area A and 100-meter area), and supports the fire department in coordinating rescue operations; The broadcasting equipment in Warehouse Area A receives instructions in the response plan document and automatically plays voice instructions, such as "People within 100 meters east of Warehouse Area A, please evacuate along the north emergency exit." The voice broadcast frequency is once every 30 seconds for 5 minutes (based on emergency broadcast guidelines).

[0046] Based on the emergency plan documents, the emergency management platform generates a visual interface that displays the risk location, impact scope, resource deployment, and execution steps, as follows: Risk location: Warehouse Area A, marked with the coordinates of the center of the chemical warehouse; Impact range: 100-meter area, with warehouse area A as the center, drawn as a circular range (based on the map data in the response plan document); Resource deployment: The location of fire trucks (e.g., adjacent fire station, 2 km away) and emergency personnel (e.g., north entrance of warehouse area A) are marked with icons; Execution steps: Display the order of command execution in the form of a timeline, for example, start: broadcast evacuation instructions; 2 minutes later: fire truck sets out; 5 minutes later: fire truck arrives; The visualization interface is deployed on managers' mobile devices and fire department command equipment, supporting real-time updates. For example, when the location of a fire truck changes, the visualization interface will update its coordinates synchronously. The visual interface integrates a semantic reasoning rule library to dynamically optimize response plans based on real-time multi-source data, as follows: Obtain real-time multi-source data from the cloud database, extract JSON structured fields, call the semantic reasoning rule library, input the JSON structured fields of the real-time multi-source data and the fields of the plan document into the semantic reasoning rule library, and generate optimization recommendations, such as "It is recommended to add one fire truck and adjust resource allocation to four vehicles"; Optimization recommendations are displayed on a visual interface for managers to confirm or modify, for example, confirming "add one fire truck" or adjusting it to "add two." The optimized plan document is stored in JSON format, uploaded to a cloud database, and redistributed to receiving terminals. The emergency management platform records all distribution and display operations and generates an operation log, which includes the sending time, receiving terminal (manager's mobile device, fire department's command equipment, broadcasting equipment in warehouse area A), adjustment records (for example, the manager adjusts resource allocation, adding 2 emergency personnel, adjusting the number from 10 to 12) and optimization recommendation records. The operation log is stored in the cloud database in JSON format.

[0047] This embodiment also provides a computer device, which is suitable for the case of a campus security risk warning and response plan generation method, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the campus security risk warning and response plan generation method proposed in the above embodiment.

[0048] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0049] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for generating a campus security risk warning and response plan as proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0050] In summary, this invention achieves accurate modeling of complex risk factors and their interrelationships within the park by constructing a knowledge graph and generating a triple structure of nodes, edges, and attributes. It not only integrates historical accident data with real-time monitoring information, but also accurately captures the causal relationships between different risk factors. It uses a semantic reasoning rule library combined with a random forest model to assess the risk type, level, and probability of these paths, improving the accuracy and reliability of risk predictions and generating detailed risk assessment reports. It not only overcomes the limitations of traditional single-dimensional analysis but also makes emergency response measures more reasonable, dynamically adjusts plans, optimizes resource allocation, and improves the speed and efficiency of responding to emergencies, ultimately achieving the goal of reducing park safety risks and protecting the safety of people and property.

[0051] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for generating a park safety risk warning and response plan, characterized by: include, Collect and pre-process multi-source data, including environmental data, personnel activity data, and equipment status data; Construct a knowledge graph to generate a triple structure of nodes, edges, and attributes. Then, integrate the triple structure and pre-processed multi-source data through path query to obtain the risk path. Use the semantic reasoning rule library and random forest model to evaluate the type, level, and probability of the risk path and generate a risk assessment report. The graph neural network model analyzes the interactive impact of multiple risk paths and obtains the interactive risk probability. Based on the interactive risk probability, preprocessed multi-source data, and priority rules, it outputs priority adjustment suggestions, allocates rescue resources, and generates an optimized plan. Distribute optimized plans to terminal devices through the emergency management platform, display the plan content using a visual interface, further optimize the plan based on the semantic reasoning rule library, and generate operation logs.

2. The method for generating a park security risk warning and response plan according to claim 1, wherein: The construction of the knowledge graph to generate a triple structure of nodes, edges and attributes refers to collecting risk information from historical accident data and safety guidelines, extracting nodes and edges from the risk information using natural language processing technology, determining the causal weights of the edges and the node risk thresholds as attributes through statistical analysis, importing the nodes, edges and attributes into a graph database, and verifying the accuracy of the node risk thresholds and edge weights by comparing historical accident data to form a structured knowledge representation that can be stored and queried.

3. The method for generating a park security risk warning and response plan according to claim 2, wherein: The node is of entity type; The edges are relationships between entity types; The attributes are the node risk threshold and the edge causality weight.

4. The method for generating a park security risk warning and response plan according to claim 3, wherein: The entity types include risk type, environmental factors, equipment status, personnel density, resources and spatial location; The edges include causal relationships, spatial relationships, and state dependency relationships.

5. The method for generating a park security risk warning and response plan according to claim 1, wherein: The method of obtaining the risk path by integrating the triple structure and multi-source data through path query refers to activating the entity types that meet the node risk threshold based on the nodes, edges, attributes and multi-source data of the knowledge graph, querying the causal relationship path connected by the edge through the path query function of the graph database, and identifying the risk path.

6. The method for generating a park security risk warning and response plan according to claim 1, wherein: The semantic reasoning rule base refers to the generation of reasoning rules based on the causal relationship of historical accident data and the safety threshold of safety criteria, comparing the values ​​of multi-source data with the reasoning rules to determine the type and level of risk path; The random forest model refers to processing the numerical values ​​of multi-source data and the attributes of the knowledge graph through multiple decision trees, counting the decision tree results and predicting the risk probability of the risk path through voting.

7. The method for generating a park security risk warning and response plan according to claim 1, wherein: The analysis of the interactive impact of multiple risk paths through a graph neural network model to obtain the interactive risk probability refers to mapping the risk path into a graph structure, utilizing node embedding and message passing mechanisms to quantify the interactions between multiple risk paths, and generate the interactive risk probability.

8. The method for generating a park security risk warning and response plan according to claim 1, wherein: The priority rule is based on historical accident data and sets a priority threshold. When the interaction risk probability exceeds the priority threshold, a high-risk interaction scenario is determined and the priority rule is triggered.

Images