Chemical facility accident hazard assessment method, device, equipment and medium

Abstract

The invention discloses a chemical facility accident hazard assessment method, device, equipment and medium, and relates to the technical field of chemical facility accident assessment, the method comprises the following steps: extracting feature dimensions of accident hazards generated by chemical facility accidents from public opinion data, and extracting feature data from the public opinion data by using a large language model; furthermore, the text feature data are numeralized, and resources are saved for subsequent calculation; under the condition that the accident hazard feature data is less, the generative adversarial network commonly used in the picture field is used for generating more accident hazard feature data, so that data enhancement and expansion are realized; and accident hazard grade division is carried out on the extracted feature data by using a clustering algorithm, so that the influence of human subjective judgment is reduced, and the manpower and time cost required by manual division is saved.

Description

Technical Field

[0001] The present invention relates to the technical field of chemical facility accident assessment, and in particular to a chemical facility accident hazard assessment method, device, equipment and medium based on large language model analysis of public opinion information. Background Art

[0002] Chemical facility accidents may cause a large number of casualties, and their harmful effects will not stop immediately after the accident. They usually last for a long time, and the scope of impact may expand with changes in the environment. Once a chemical facility accident occurs, it is easy to cause casualties among the surrounding people and cause long-term damage to the environment and ecosystem. Accurate assessment of the hazards after a chemical facility accident will help the government formulate effective protective measures to control the impact of the disaster.

[0003] In the Internet era, the spread of public opinion information is often faster than that of official information, and the information sources are also more extensive. Analyzing public opinion information can help us quickly and comprehensively grasp the trends of events.

[0004] However, unlike professional and authoritative data, the authenticity of public opinion data needs to be carefully identified, and the correlation between data also needs to be fully considered. The impact of chemical facility accident hazards is the result of long-term dynamic monitoring, which requires long-term analysis of multi-faceted data in order to provide decision makers with a reliable reference.

[0005] Deep learning is widely used in processing public opinion text information, and compared with other deep learning models, large language models have great advantages, especially in text analysis, important information extraction, and text data generation.

[0006] Therefore, how to provide a method for evaluating the hazards of chemical facility accidents based on analyzing public opinion information with a large language model is a technical problem that urgently needs to be solved by those skilled in the art. Summary of the invention

[0007] In view of the above problems, the present invention provides a chemical facility accident hazard assessment method, device, equipment and medium for overcoming the above problems or at least partially solving the above problems.

[0008] The present invention provides the following scheme:

[0009] A chemical facility accident hazard assessment method, comprising:

[0010] Obtaining public opinion data to be processed, wherein the public opinion data to be processed includes chemical facility accident data to be processed;

[0011] Extracting feature data of the chemical facility accident data to be processed using the trained large language model, wherein the feature dimensions of the feature data include influencing factors that affect the occurrence of accident hazards;

[0012] Using the trained classification model in combination with the characteristic data to perform identification and obtain an assessment result of the accident hazard;

[0013] The large language model training data is obtained in the following manner:

[0014] Collect historical chemical facility accident data that occurred in the past by matching key information;

[0015] The historical chemical facility accident data is enhanced and expanded through a generative adversarial network;

[0016] The classification model training data is obtained in the following way:

[0017] The characteristic data is classified into accident hazard levels by using a clustering algorithm to obtain an accident hazard level;

[0018] The accident hazard level is added to the feature data to obtain supervised training data.

[0019] Preferably: the influencing factors at least include whether there are residential buildings nearby, economic losses, new deaths, new injuries, whether it is located in a densely populated area, whether nearby residential buildings are affected, whether it affects nearby water sources, whether it affects the atmosphere, and whether it affects the soil.

[0020] Preferably: the generative adversarial network includes a generator and a discriminator; the hidden layer of the generator includes any one of a transposed convolutional layer and a fully connected layer; the discriminator includes any one of a convolutional neural network and a fully connected neural network;

[0021] The generator and the discriminator are made to reach Nash equilibrium, thereby obtaining extended data with the same distribution as the original data, and completing the data enhancement operation.

[0022] Preferably: the clustering algorithm is configured to classify the results obtained through clustering analysis according to the data characterization, and add the level of each feature data to the last column of the corresponding feature list data; the level is defined according to the actual severity represented by the dimension.

[0023] Preferably, the clustering algorithm includes any one of K-means, hierarchical clustering, DBSCAN, and Gaussian mixture model.

[0024] Preferably, the classification model is constructed based on any one of random forest, artificial neural network and convolutional neural network.

[0025] Preferably, the historical chemical facility accident data includes public opinion data related to historical chemical plant explosion accidents, which are sorted in chronological order of accident development.

[0026] A chemical facility accident hazard assessment device, used to perform the above-mentioned chemical facility accident hazard assessment method, the device comprising:

[0027] A data acquisition unit, used to acquire public opinion data to be processed, wherein the public opinion data to be processed includes chemical facility accident data to be processed;

[0028] A feature extraction unit, used to extract feature data of the chemical facility accident data to be processed using the trained large language model, wherein the feature dimensions of the feature data include influencing factors that affect the occurrence of accident hazards;

[0029] An evaluation unit, used to use the trained classification model in combination with the feature data to perform identification and obtain an evaluation result of the accident hazard;

[0030] The large language model training data is obtained in the following manner:

[0031] Collect historical chemical facility accident data that occurred in the past by matching key information;

[0032] The historical chemical facility accident data is enhanced and expanded through a generative adversarial network;

[0033] The classification model training data is obtained in the following way:

[0034] The characteristic data is classified into accident hazard levels by using a clustering algorithm to obtain an accident hazard level;

[0035] The accident hazard level is added to the feature data to obtain supervised training data.

[0036] A chemical facility accident hazard assessment device, the device comprising a processor and a memory:

[0037] The memory is used to store program codes and transmit the program codes to the processor;

[0038] The processor is used to execute the above-mentioned chemical facility accident hazard assessment method according to the instructions in the program code.

[0039] A computer-readable storage medium is used to store program codes, and the program codes are used to execute the above-mentioned chemical facility accident hazard assessment method.

[0040] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0041] The embodiments of the present application provide a method, device, equipment and medium for assessing the hazards of chemical facility accidents. The method extracts characteristic dimensions of accident hazards caused by chemical facility accidents from public opinion data, and uses a large language model to extract characteristic data from the public opinion data, thereby digitizing the text characteristic data to save resources for subsequent calculations. When there is less accident hazard characteristic data, a generative adversarial network commonly used in the image field is used to generate more accident hazard characteristic data, thereby achieving data enhancement and expansion. A clustering algorithm is used to classify the extracted characteristic data into accident hazard levels, thereby reducing the impact of human subjective judgment and saving the manpower and time costs required for manual classification.

[0042] Of course, any product implementing the present invention does not necessarily need to achieve all of the above-mentioned advantages at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0044] Figure 1 is a flow chart of a chemical facility accident hazard assessment method provided by an embodiment of the present invention;

[0045] Figure 2 is a schematic diagram of a model framework provided by an embodiment of the present invention;

[0046] Figure 3 It is a schematic diagram of data collection analysis and feature dimension extraction provided by an embodiment of the present invention;

[0047] Figure 4 is a schematic diagram of data enhancement provided by an embodiment of the present invention;

[0048] Figure 5 It is a schematic diagram of accident hazard level classification and assessment provided by an embodiment of the present invention;

[0049] Figure 6 is a schematic diagram of a chemical facility accident hazard assessment device provided by an embodiment of the present invention;

[0050] Figure 7 It is a schematic diagram of a chemical facility accident hazard assessment device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0051] The technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all of the embodiments. Based on the embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0052] See also Figure 1 , is a chemical facility accident hazard assessment method provided by an embodiment of the present invention, such as Figure 1 As shown, the method may include:

[0053] S101: Acquire public opinion data to be processed, where the public opinion data to be processed includes chemical facility accident data to be processed;

[0054] S102: Use the trained large language model to extract feature data of the chemical facility accident data to be processed, the feature dimensions of the feature data include influencing factors that affect the occurrence of accident hazards; in specific implementation, the embodiment of the present application can also provide that the influencing factors at least include whether there are residential buildings nearby, economic losses, new deaths, new injuries, whether it is located in a densely populated area, whether nearby residential buildings are affected, whether it affects nearby water sources, whether it affects the atmosphere, and whether it affects the soil.

[0055] S103: using the trained classification model in combination with the feature data to perform identification and obtain an assessment result of the accident hazard;

[0056] The large language model training data is obtained in the following manner:

[0057] The historical chemical facility accident data that occurred in the past are collected by matching key information; the historical chemical facility accident data include public opinion data related to historical chemical plant explosion accidents, which are sorted in chronological order according to the development of the accidents.

[0058] The historical chemical facility accident data is enhanced and expanded by a generative adversarial network; in specific implementation, the embodiment of the present application can provide that the generative adversarial network includes a generator and a discriminator; the hidden layer of the generator includes any one of a transposed convolutional layer and a fully connected layer; the discriminator includes any one of a convolutional neural network and a fully connected neural network;

[0059] The generator and the discriminator are made to reach Nash equilibrium, thereby obtaining extended data with the same distribution as the original data, and completing the data enhancement operation.

[0060] The classification model training data is obtained in the following way:

[0061] The feature data is classified into accident hazard levels through a clustering algorithm to obtain the accident hazard level; in specific implementation, the embodiment of the present application can provide that the clustering algorithm is configured to classify the results obtained through clustering analysis according to the data characterization, and the level of each feature data is added to the last column of the corresponding feature list data; the level is defined according to the actual severity represented by the dimension.

[0062] Furthermore, the clustering algorithm includes any one of K-means, hierarchical clustering, DBSCAN, and Gaussian mixture model. The classification model is built based on any one of random forest, artificial neural network, and convolutional neural network.

[0063] The accident hazard level is added to the feature data to obtain supervised training data.

[0064] The chemical facility accident hazard assessment method provided in the embodiment of the present application conducts an assessment of the chemical facility accident hazard based on open source public opinion data. First, it is necessary to collect historical data of specified chemical facility accidents, but such data is relatively small and data expansion is required. Then, based on the collected data, analysis is performed to extract factors that affect the occurrence of accident hazards. These factors are used as characteristic dimensions of the accident hazards, and features are extracted through a large language model to obtain characteristic data of the event or accident. The obtained characteristic data is divided into categories through a clustering method to obtain the level of the accident hazard, and the obtained level is added as a label to the characteristic data, and complete supervised training data is obtained at this time. Build a classification model, use the final data to train the classification model, obtain the final classification model, and realize the assessment function of the accident hazard.

[0065] The following is a detailed description of the chemical facility accident hazard assessment method provided in the embodiments of the present application. Figure 2 shown.

[0066] 1. Data collection.

[0067] First, the data of chemical facility accidents that have occurred in history are collected by matching key information. This application is for analysis of public opinion data, so the collected data is public opinion data disclosed on the Internet. There are many factors that affect the occurrence of chemical facility accident hazards, which requires analysis based on chemical facility accident events that have occurred in history. Since the amount of data on chemical facility accident events in history is relatively small, and they are mainly chemical plant explosion accidents, the collected data are mainly public opinion data related to chemical plant explosion accidents.

[0068] (1) When collecting data, we mainly focus on historical chemical facility accidents, and organize them in chronological order to obtain the development context of each historical event.

[0069] (2) Since there are relatively few chemical facility accidents in history, it is not conducive to effective training models. After analysis, chemical plant accidents are the main chemical facility accidents. Therefore, this application collects public opinion data on chemical plant accidents in history. The collection method is to organize them in chronological order according to the development of the accidents.

[0070] 2. Data analysis and feature dimension extraction.

[0071] Analyze the corresponding academic literature and public opinion data, obtain the influencing factors in the development of each event or accident, and organize these factors into characteristic information that affects the hazards of the accident; analyze the development context of each event based on the collected public opinion data, and the content of the analysis is mainly the influencing factors of the accident hazards.

[0072] (1) The accident hazards of each event have special influencing factors. The present invention will extract common influencing factors, such as: whether there are residential buildings nearby, economic losses (10,000 yuan), the number of new deaths, the number of new injuries, whether it is located in a densely populated area, whether nearby residential buildings are affected, whether it affects nearby water sources, whether it affects the atmosphere, whether it affects the soil, etc.

[0073] (2) Use the large language model to standardize and extract the features contained in the public opinion information, so as to obtain the list data representing the accident hazards. In the specific implementation, according to the analyzed feature data, the large language model is used to fuse the collected public opinions and extract the features, so as to obtain the list data representing the accident hazards. The schematic diagram of data collection analysis and feature dimension extraction is shown in the figure. Figure 3 shown.

[0074] 3. Data Augmentation

[0075] Since the amount of data on chemical facility accidents is relatively small, the data is enhanced and expanded through generative adversarial networks. In the data collection and analysis stage, even if the data source is expanded, the total amount of data is still relatively small. At this time, the data is enhanced through generative adversarial networks.

[0076] (1) Build a generative adversarial network and select appropriate networks as generators and discriminators based on the characteristics of the data. When building the hidden layer of the generator, use including but not limited to transposed convolutional layers and fully connected layers; when building the discriminator, use including but not limited to convolutional neural networks and fully connected neural networks.

[0077] (2) Train the network and use the feature data to train the generative adversarial network so that the generator and the discriminator reach Nash equilibrium, thereby obtaining extended data with the same distribution as the original data and completing the data enhancement operation. Figure 4 shown.

[0078] 4. Classification of accident hazard levels

[0079] There is no clear national or international standard for the level of accident hazards. In order to confirm the level of accident hazards, the obtained data is clustered and divided. In specific implementation, since there is no national or international standard for accident hazards, if artificial division is performed, authoritative experts are required to conduct a comprehensive analysis and judgment of the incident, which often requires a lot of manpower and time costs. Since there is a certain correlation and similarity within the feature data, it can be divided according to the characteristics of data distribution. Clustering algorithms have great advantages in this regard.

[0080] (1) Select an appropriate clustering algorithm to divide the data, and define the levels of the divided categories according to the actual severity represented by the dimension. Use a clustering algorithm to perform cluster analysis on the obtained data. The clustering algorithm that can be used includes but is not limited to: K-means, hierarchical clustering, DBSCAN, and Gaussian mixture model (GMM).

[0081] (2) The results obtained through cluster analysis are classified into levels according to the data representation, and the severity is divided from high to low, with level 1 being the most serious level, and then ranked in order.

[0082] (3) Add the level of each data to the last column of the corresponding feature list data to facilitate training of the classification model.

[0083] After the above processing, the characteristic information representing the accident hazard and the corresponding accident hazard level can be obtained. In order to accurately assess the level of accident hazard, a classification model is built to achieve this.

[0084] 5. Accident hazard assessment.

[0085] Build a classification model and use the data obtained above to train the classification model.

[0086] (1) Building a classification model. The classification model built by the present invention includes but is not limited to: random forest, fully connected neural network, and convolutional neural network.

[0087] (2) Divide the acquired data into a training set and a test set, train the model, and obtain a trained model.

[0088] (3) The newly collected public opinion data to be processed is subjected to information fusion and feature extraction using a large language model to obtain feature data that characterizes the hazard level of the incident. The feature data is then identified using a trained classification model to ultimately obtain an assessment of the hazard level of the incident. Figure 5 shown.

[0089] In summary, the chemical facility accident hazard assessment method provided in the present application extracts the characteristic dimensions of accident hazards caused by chemical facility accidents from public opinion data, and uses a large language model to extract feature data from the public opinion data, and then digitizes the text feature data to save resources for subsequent calculations; when there is less accident hazard feature data, the generative adversarial network commonly used in the image field is used to generate more accident hazard feature data, thereby achieving data enhancement and expansion; a clustering algorithm is used to classify the extracted feature data into accident hazard levels, thereby reducing the impact of human subjective judgment and saving the manpower and time costs required for manual classification.

[0090] See also Figure 6 , the embodiment of the present application can also provide a chemical facility accident hazard assessment device, such as Figure 6 As shown, the device for performing the above-mentioned chemical facility accident hazard assessment method may include:

[0091] The data acquisition unit 601 is used to acquire public opinion data to be processed, wherein the public opinion data to be processed includes chemical facility accident data to be processed;

[0092] A feature extraction unit 602 is used to extract feature data of the chemical facility accident data to be processed using the trained large language model, wherein the feature dimensions of the feature data include influencing factors that affect the occurrence of accident hazards;

[0093] An evaluation unit 603 is used to use the trained classification model in combination with the feature data to perform identification and obtain an evaluation result of the accident hazard;

[0094] The large language model training data is obtained in the following manner:

[0095] Collect historical chemical facility accident data that occurred in the past by matching key information;

[0096] The historical chemical facility accident data is enhanced and expanded through a generative adversarial network;

[0097] The classification model training data is obtained in the following way:

[0098] The characteristic data is classified into accident hazard levels by using a clustering algorithm to obtain an accident hazard level;

[0099] The accident hazard level is added to the feature data to obtain supervised training data.

[0100] The embodiment of the present application may also provide a chemical facility accident hazard assessment device, the device comprising a processor and a memory:

[0101] The memory is used to store program codes and transmit the program codes to the processor;

[0102] The processor is used to execute the steps of the above-mentioned chemical facility accident hazard assessment method according to the instructions in the program code.

[0103] like Figure 7 As shown, a chemical facility accident hazard assessment device provided by an embodiment of the present application may include: a processor 10, a memory 11, a communication interface 12 and a communication bus 13. The processor 10, the memory 11 and the communication interface 12 all communicate with each other through the communication bus 13.

[0104] In the embodiment of the present application, the processor 10 may be a central processing unit (CPU), an application specific integrated circuit, a digital signal processor, a field programmable gate array or other programmable logic devices.

[0105] The processor 10 may call a program stored in the memory 11. Specifically, the processor 10 may execute operations in an embodiment of a chemical facility accident hazard assessment method.

[0106] The memory 11 is used to store one or more programs, which may include program codes, and the program codes include computer operation instructions. In the embodiment of the present application, the memory 11 at least stores programs for implementing the following functions:

[0107] Obtaining public opinion data to be processed, wherein the public opinion data to be processed includes chemical facility accident data to be processed;

[0108] Extracting feature data of the chemical facility accident data to be processed using the trained large language model, wherein the feature dimensions of the feature data include influencing factors that affect the occurrence of accident hazards;

[0109] Using the trained classification model in combination with the characteristic data to perform identification and obtain an assessment result of the accident hazard;

[0110] The large language model training data is obtained in the following manner:

[0111] Collect historical chemical facility accident data that occurred in the past by matching key information;

[0112] The historical chemical facility accident data is enhanced and expanded through a generative adversarial network;

[0113] The classification model training data is obtained in the following way:

[0114] The characteristic data is classified into accident hazard levels by using a clustering algorithm to obtain an accident hazard level;

[0115] The accident hazard level is added to the feature data to obtain supervised training data.

[0116] In one possible implementation, the memory 11 may include a program storage area and a data storage area, wherein the program storage area can store an operating system and application programs required for at least one function (such as a file creation function, a data reading and writing function), etc.; the data storage area can store data created during use, such as initialization data, etc.

[0117] In addition, the memory 11 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device or other volatile solid-state storage device.

[0118] The communication interface 12 may be an interface of a communication module, and is used to connect to other devices or systems.

[0119] Of course, it should be noted that Figure 7 The structure shown does not constitute a limitation on the chemical facility accident hazard assessment device in the embodiment of the present application. In actual applications, the chemical facility accident hazard assessment device may include Figure 7 More or fewer components than shown, or combinations of certain components.

[0120] The embodiment of the present application may also provide a computer-readable storage medium, wherein the computer-readable storage medium is used to store program code, and the program code is used to execute the steps of the above-mentioned chemical facility accident hazard assessment method.

[0121] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0122] It can be known from the description of the above implementation methods that those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present application or certain parts of the embodiments.

[0123] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The system and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0124] The above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention are included in the protection scope of the present invention.

Claims

1. A method for assessing the hazard of a chemical facility accident, characterized in that: include: Obtaining public opinion data to be processed, wherein the public opinion data to be processed includes chemical facility accident data to be processed; Extracting feature data of the chemical facility accident data to be processed using the trained large language model, wherein the feature dimensions of the feature data include influencing factors that affect the occurrence of accident hazards; Using the trained classification model in combination with the characteristic data to perform identification and obtain an assessment result of the accident hazard; The large language model training data is obtained in the following manner: Collect historical chemical facility accident data that occurred in the past by matching key information; The historical chemical facility accident data is enhanced and expanded through a generative adversarial network; The classification model training data is obtained in the following way: The characteristic data is classified into accident hazard levels by using a clustering algorithm to obtain an accident hazard level; The accident hazard level is added to the feature data to obtain supervised training data.

2. The chemical facility accident hazard assessment method according to claim 1, characterized in that: The influencing factors include at least whether there are residential houses nearby, economic losses, new deaths, new injuries, whether it is located in a densely populated area, whether nearby residential buildings are affected, whether it affects nearby water sources, whether it affects the atmosphere, and whether it affects the soil.

3. The chemical facility accident hazard assessment method according to claim 1, characterized in that: The generative adversarial network includes a generator and a discriminator; the hidden layer of the generator includes any one of a transposed convolutional layer and a fully connected layer; the discriminator includes any one of a convolutional neural network and a fully connected neural network; The generator and the discriminator are made to reach Nash equilibrium, thereby obtaining extended data with the same distribution as the original data, and completing the data enhancement operation.

4. The chemical facility accident hazard assessment method according to claim 1, characterized in that: The clustering algorithm is configured to classify the results obtained through clustering analysis according to the data representation, and add the level of each feature data to the last column of the corresponding feature list data; the level is defined according to the actual severity represented by the dimension.

5. The chemical facility accident hazard assessment method according to claim 1, characterized in that: The clustering algorithm includes any one of K-means, hierarchical clustering, DBSCAN, and Gaussian mixture model.

6. The chemical facility accident hazard assessment method according to claim 1, characterized in that: The classification model is constructed based on any one of random forest, artificial neural network and convolutional neural network.

7. The chemical facility accident hazard assessment method according to claim 1, characterized in that: The historical chemical facility accident data include public opinion data related to historical chemical plant explosion accidents, which are sorted in chronological order according to the development of the accidents.

8. A chemical facility accident hazard assessment device, characterized in that: Used to implement the chemical facility accident hazard assessment method according to any one of claims 1 to 7, the device comprises: A data acquisition unit, used to acquire public opinion data to be processed, wherein the public opinion data to be processed includes chemical facility accident data to be processed; A feature extraction unit, used to extract feature data of the chemical facility accident data to be processed using the trained large language model, wherein the feature dimensions of the feature data include influencing factors that affect the occurrence of accident hazards; An evaluation unit, used to use the trained classification model in combination with the feature data to perform identification and obtain an evaluation result of the accident hazard; The large language model training data is obtained in the following manner: Collect historical chemical facility accident data that occurred in the past by matching key information; The historical chemical facility accident data is enhanced and expanded through a generative adversarial network; The classification model training data is obtained in the following way: The characteristic data is classified into accident hazard levels by using a clustering algorithm to obtain an accident hazard level; The accident hazard level is added to the feature data to obtain supervised training data.

9. A chemical facility accident hazard assessment device, characterized in that: The device comprises a processor and a memory: The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the chemical facility accident hazard assessment method according to any one of claims 1 to 7 according to the instructions in the program code.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store program codes, and the program codes are used to execute the chemical facility accident hazard assessment method according to any one of claims 1-7.

Images