Emergency response decision optimization system and method for intelligent refrigerant system

Abstract

The invention discloses an emergency response decision optimization system and method for an intelligent refrigerant system, and the method comprises the steps: carrying out the automatic data collection, crawling and collecting the text data of an emergency plan, a historical emergency event and a risk factor related to the refrigerant system, carrying out the intelligent text preprocessing, and carrying out the decision optimization. Potential association relationships in the data are mined through an association rule mining algorithm, and a multi-layer complex network model is constructed. The multi-layer complex network model is input to a convolutional neural network through an embedded layer, key features are automatically extracted, and the model is trained, so that optimal emergency response strategies in different risk scenes are learned. The system receives risk input data in real time through an application program interface, and rapidly outputs optimized emergency resource configuration and coping measure combination to assist a decision maker to make an optimal decision in an emergency.

Description

Technical Field

[0001] The present invention relates to the technical field of emergency response decision matching for refrigerant systems, and particularly to an emergency plan decision optimization system for refrigerant systems based on artificial intelligence and complex networks and its implementation method. Background Art

[0002] Refrigerant systems are an important component widely used in modern refrigeration and air conditioning systems. Any failure may have a significant impact on the normal operation of the system and even cause serious environmental and safety problems. Traditional emergency responses for refrigerant systems mainly rely on predefined rules or the experience of operators. However, in the event of an emergency in a refrigerant system, such static emergency plans are difficult to quickly and effectively provide targeted solutions, cannot dynamically adjust response measures according to the actual situation, and cannot achieve efficient scheduling of emergency resources. This traditional method is inefficient and has a lag in the decision-making process when dealing with complex emergencies.

[0003] With the development of complex network theory, artificial intelligence, and big data technology, optimizing emergency decision-making based on these intelligent means has become an important research direction in the safety management of refrigerant systems. Complex network theory can effectively describe the correlation relationships between the various components in a refrigerant system, helping to identify key nodes and resources, while machine learning algorithms (convolutional neural networks) can learn the potential correlations between emergency plans, failure modes, risk factors, and resource requirements from historical data, realizing an intelligent emergency response decision support system. Such an emergency response optimization system based on complex networks and artificial intelligence can dynamically adjust response measures in the event of an emergency in a refrigerant system, provide fast and accurate emergency resource scheduling and response strategies, and effectively improve the emergency handling ability and response speed of the system. Summary of the Invention

[0004] The present invention aims to provide an emergency response decision optimization system for refrigerant systems based on artificial intelligence and complex networks and its implementation method to solve at least one of the technical problems existing in the above background art. The system aims to handle emergencies such as equipment failures, pipeline ruptures, and leaks that may occur in a refrigerant system, and provide the best emergency response plan through intelligent means to ensure the safe and efficient operation of the system.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] On the one hand, the present invention provides an emergency response decision optimization system for refrigerant systems based on artificial intelligence and complex networks, including the following modules:

[0007] Data Collection Module: Automatically collect emergency plan data, historical emergency event data, and risk factor data related to the refrigerant system through web crawler technology. The data sources are extensive, covering government emergency management systems, industry databases, and authoritative news media, etc.;

[0008] Data Preprocessing Module: Used to preprocess the collected text data. The specific steps include removing stop words, text tokenization, and extracting keywords in the text using natural language processing (NLP) technology. The extraction of keywords covers risk factors, emergency resources, and response measures, etc.;

[0009] Association Rule Mining Module: Based on the association rule mining algorithm, deeply analyze the structured data. By screening support, confidence, and lift, identify strong association rules with high confidence for the construction of subsequent multi-layer complex network models;

[0010] Complex Network Construction Module: Based on the mined association rules, construct a multi-layer complex network model. The model consists of three layers of nodes: the first layer of nodes represents risk factors, the second layer of nodes represents emergency resources, and the third layer of nodes represents response measures. The connections between nodes represent causal relationships or associations of resource requirements;

[0011] Machine Learning Training Module: Adopt a convolutional neural network, take the multi-layer complex network model as input, and map the associations between risk factors, emergency resources, and response measures to a high-dimensional feature space through deep learning. The model outputs the optimal combination of emergency resources and response measures for specific risk scenarios to form an emergency response decision;

[0012] Application Programming Interface (API): Provide a user interaction platform to support real-time input of risk information. The system automatically matches based on the input information and outputs the optimal combination of emergency resources and response measures. At the same time, it supports dynamic data updates, enabling the system to adjust decisions according to real-time situations and continuously optimize the emergency response strategy.

[0013] Furthermore, it also includes: Model Optimization Module, which integrates multi-objective optimization algorithms and can dynamically optimize the emergency response plan according to different emergency scenarios to meet the requirements of multiple preset optimization goals.

[0014] The Model Optimization Module dynamically adjusts the model parameters through the real-time feedback of emergency event data to enhance the system's response ability and prediction accuracy in different risk scenarios.

[0015] Furthermore, it also includes: System Expansion Module, which automatically expands the nodes and edges in the complex network model as the historical emergency event data and risk factor data accumulate.

[0016] The complex network model construction module constructs a three-layer network according to the association rules to represent the relationships between entities. The first layer is the risk evolution network layer, where the nodes represent risk factors and the edges represent the evolution paths and causal relationships between risk factors; the second layer is the emergency resource network layer, where the nodes represent emergency resources and the edges represent the associations between resources; the third layer is the emergency plan network layer, where the nodes represent response measures and the edges represent the process associations between measures.

[0017] By analyzing the emergency resource requirements and emergency measures corresponding to a certain risk factor in historical emergency events, the system organically connects the constructed risk evolution network layer, emergency resource network layer, and emergency plan network layer to form a multi-layer complex network model.

[0018] Each node represents a specific entity, such as a risk factor, an emergency resource, or a response measure. Each node is assigned a unique serial number, and all the entities represented by the serial numbers are stored in the document. The constructed complex network is represented in the form of node pairs and input into the convolutional neural network training model. This input no longer uses text or word embeddings, but is input through the network structure represented by node serial numbers and node pairs, enabling the model to focus more on learning the features of the network structure.

[0019] The system extension module ensures the uniqueness of nodes during system extension and upgrade through automatic assignment of unique serial numbers and achieves seamless connection.

[0020] The machine learning training module contains an input layer for receiving the constructed multi-layer complex network model. The model matches specific risk factors with emergency resources and response measures by learning the network structure in historical data.

[0021] The convolutional neural network is used to process the input network model to learn the relationship patterns between nodes. The convolutional layer processes node serial numbers and corresponding relationships, rather than text information.

[0022] The input network model is divided into a training set, a validation set, and a test set to ensure the scientific nature of model training and performance evaluation.

[0023] During model training, positive and negative samples are divided. Positive samples are the resources and measures used in historical events, and negative samples are the resources and measures not used, to improve the prediction accuracy of the model.

[0024] By distinguishing between positive and negative samples, the model can not only predict the resources and measures to be used in a given accident scenario, but also avoid predicting unnecessary or inappropriate resources and measures.

[0025] During the training process, the model is evaluated and optimized through accuracy, recall rate, and F1 score to ensure that the model can provide accurate emergency response decisions in new accident scenarios.

[0026] The application programming interface (API) is used to receive the input accident risk information, map it to the corresponding node numbers, and output the optimal combination of emergency resources and response measures according to the trained convolutional neural network model, providing the best emergency response suggestions for decision-makers.

[0027] In a second aspect, the present invention provides an emergency response decision optimization system for a refrigerant system based on artificial intelligence and complex networks, including:

[0028] Step 1: Through the data collection module, automatically crawl historical emergency event data, accident cases, and emergency plan texts related to the refrigerant system from the government emergency management platform, industry databases, and authoritative news media. The collected data covers detailed records of historical events, accident descriptions, and the implementation processes of emergency plans. To ensure the authority and comprehensiveness of the data, priority is given to content from official government channels, industry standard databases, and certified authoritative media. The crawled data undergoes preliminary screening and cleaning to remove redundant and irrelevant information, ensuring the accuracy and applicability of the data and laying a foundation for subsequent keyword extraction and correlation analysis.

[0029] Step 2: Preprocess the cleaned text data. Use the TextRank algorithm to filter stop words and perform word segmentation on the text data to extract core keywords. In specific operations, extract emergency resource requirements from historical emergency event texts, extract risk factors from accident cases, and identify response measures from emergency plan texts. The extracted keyword information provides data support for subsequent construction of the complex network model.

[0030] Step 3: Use the weighted Eclat algorithm to mine association rules from the processed text data and analyze the association relationships between keywords. By calculating indicators such as support, confidence, and lift, extract strong association rules and construct a complex network model based on them. The keywords serve as nodes in the network, and the association relationships between nodes serve as edges in the network, forming a complete network structure.

[0031] Step 4: According to the mining results of the association rules, construct a three-layer complex network model: The first layer is the risk evolution network layer, where nodes represent risk factors and edges represent risk evolution paths and causal relationships; the second layer is the emergency resource network layer, where nodes represent the emergency resources required in specific risk scenarios and edges are constructed based on the correlation between resources; the third layer is the emergency plan network layer, where nodes represent the specific response measures in the emergency plan and edges are constructed according to the logical relationships between the response measures. Through the association relationships among risk factors, emergency resources, and response measures, the three layers of the network are organically integrated to form a multi-level complex network model.

[0032] Step 5: Each node corresponds to a unique entity, such as a risk factor, emergency resource, or response measure. The system assigns a unique serial number to each node and stores its specific meaning in an independent document. The constructed complex network model is input into a convolutional neural network in the form of node pairs for training, thereby simplifying data input, avoiding the complexity of text embedding or word embedding, and focusing on the learning of network structure features.

[0033] Step 6: Input the three-layer network model into the convolutional neural network in the form of node pairs. Use the stratified sampling method to divide the dataset into a training set, a validation set, and a test set to ensure the balance of edge label distribution. The training set accounts for 70%-80% of the total data volume and is used for the initial training of the model; the validation set accounts for 10%-15% and is used for hyperparameter tuning and model optimization; the test set accounts for 10%-15% and is used for the evaluation of the final performance of the model.

[0034] Step 7: During the model training process, positive and negative samples are input for optimization. Positive samples are the resources and response measures actually used in historical emergency events, while negative samples are the unused resources and measures. By distinguishing between positive and negative samples, the prediction accuracy of the model is improved, ensuring that the combination of output emergency resources and response measures is the most appropriate, and avoiding redundant resource scheduling. The system focuses on learning which resources and measures can effectively contain risk evolution.

[0035] Step 8: After training is completed, decision-making optimization is carried out in combination with a multi-objective optimization algorithm. The optimization objectives include response speed, resource utilization efficiency, and cost control. Through the Pareto optimal solution set, multiple emergency response plans are generated for decision-makers to choose from.

[0036] Step 9: Evaluate the model performance using metrics such as accuracy, recall, and F1-score, and adjust the model structure based on the evaluation results. Further improve the optimization effect of the model by increasing the data volume or introducing more features.

[0037] Step 10: Design the system application programming interface (API) to receive newly input risk data and map it to the corresponding node serial number. Based on the trained convolutional neural network model, the system outputs the corresponding emergency resources and response measures, providing the best emergency response suggestions for decision-makers.

[0038] Step 11: Continuously optimize the model parameters according to the feedback of actual emergency events to improve the system adaptability. At the same time, as more historical data accumulates, expand the network nodes, ensure the uniqueness of node serial numbers, avoid conflicts, and ensure the sustainable expansion of the system.

[0039] Advantages of the present invention: By introducing the architectures of artificial intelligence and complex networks, the limitations of traditional emergency plans relying on manual experience and fixed models are solved. In this system, through real-time data analysis and dynamic model optimization, auxiliary information based on scientific evidence can be provided to decision-makers in emergency situations, helping them make the optimal emergency decisions quickly. Compared with traditional methods, the present invention significantly improves the efficiency of emergency response and the accuracy of decision-making, reduces the interference of human factors on the decision-making process, and enhances the scientific and intelligent level of emergency management. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for description in the embodiments will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0041] Figure 1 It is a functional framework diagram of the refrigerant system emergency response decision optimization system based on artificial intelligence and complex networks according to the embodiments of the present invention.

[0042] Figure 2 It is a flowchart for generating the refrigerant system emergency response decision according to the embodiments of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] The embodiments of the present invention will be described in detail in the following drawings. In the drawings, the same or similar reference numerals are always used to represent the same or functionally similar elements. The described embodiments are only used to explain the technical solutions of the present invention and cannot be used as a limitation to the present invention.

[0044] Those skilled in the art should understand that unless otherwise clearly defined, the terms (including technical terms and scientific terms) used herein should have the meanings generally accepted by those of ordinary skill in the technical field to which they belong.

[0045] In addition, the interpretation of the terms should be consistent with their usual meanings in the prior art unless otherwise defined herein. It should not be interpreted idealistically or overly formally.

[0046] The singular forms such as "a", "an", "the" and other words used herein should be understood to cover the plural forms as well, unless the context clearly excludes them. Further, the term "including" should be interpreted as not excluding the existence of other unmentioned features, steps, operations, elements or components.

[0047] When referring to terms such as "one embodiment", "some embodiments" or "examples", the specific features, structures, materials or characteristics described can be applied to multiple embodiments, and these features in different embodiments can be flexibly combined according to actual needs.

[0048] To better understand the present invention, specific embodiments will be described in detail below with reference to the accompanying drawings. These embodiments are for illustrative purposes only and do not constitute a limitation on the scope of the present invention.

[0049] Those skilled in the art should understand that the drawings are only schematic diagrams, and the components shown are not necessary for implementing the present invention.

[0050] Figure 1 The schematic diagram of the functional modules of the emergency response decision optimization system for the refrigerant system of the present invention is shown. The system includes six core modules: a data collection module, a data preprocessing module, an association relationship mining module, a complex network construction module, a convolutional neural network module, and an application programming interface module. The historical risk events and emergency strategies of the refrigerant system are analyzed through a complex network, and the convolutional neural network is used to learn the strategies to evaluate risks in real time and output the optimal emergency response plan.

[0051] (1) Convolutional neural network for emergency response decision of refrigerant system: As the core and brain of strategy learning, the convolutional neural network can understand the strategies in historical emergency events. The role of this module is to serve as the center of the emergency response decision optimization system for the refrigerant system, receive the information output by other functional modules, and output emergency response strategies in real time according to the process of strategy learning.

[0052] (2) Data collection module: Automatically obtain text data related to emergency response plans, historical events, and risk factors of the refrigerant system from government emergency management systems, industry databases, and authoritative media through web crawlers.

[0053] (3) Data preprocessing module: Preprocess the collected text data, apply natural language processing techniques (such as Text Rank) to extract keywords, and the extracted content includes risk factors, emergency resources, and response measures, which are input as node factors into the subsequent modules.

[0054] (4) Association relationship mining module: Use the weighted Eclat algorithm to mine strong association rules in the text, screen out high-confidence associations based on indicators such as support and confidence, and use them to construct a multi-layer complex network model;

[0055] (5) Complex network construction module: Construct a multi-layer risk network model based on the extracted node factors and association relationships. This model includes three layers of nodes: the first layer is risk factors, the second layer is emergency resources, and the third layer is response measures. The edges represent the causal relationships or resource demand associations between nodes;

[0056] (6) Multi-objective optimization: The data set is divided into a training set, a validation set, and a test set using the stratified sampling method to ensure the balance of the edge label distribution. By distinguishing between positive and negative samples, the prediction accuracy of the model is improved to ensure that the combined emergency resources and response measures output are the most appropriate, and redundant resource scheduling is avoided. After training, decision optimization is performed in combination with a multi-objective optimization algorithm. The optimization objectives include response speed, resource utilization efficiency, and cost control;

[0057] (7) Application programming interface: Provide a user interaction platform to support real-time input of accident risk information and output the optimal combination of emergency resources and response measures. At the same time, it supports dynamic data updates and real-time adjustment of decision-making plans.

[0058] Figure 2 It is a schematic flowchart of the implementation method of an emergency response decision optimization system for a refrigerant system based on artificial intelligence and complex networks. Using the above system, the generation process of the emergency response strategy for the refrigerant system is realized, including four links: emergency event collection, emergency network construction, emergency strategy learning, and emergency response decision generation, which are refined into 11 steps. Among them, steps 1 and 2 belong to the emergency event collection link, steps 3, 4, and 5 belong to the emergency network construction link, steps 6, 7, 8, and 9 belong to the emergency strategy learning link, and steps 10 and 11 belong to the emergency response decision generation link.

[0059] Step 1: Use the data collection module to crawl historical emergency event data, accident cases, and emergency plan text data related to the refrigerant system from government emergency management websites, industry databases, and authoritative news media. The obtained data covers records of historical emergency events, detailed descriptions of related accidents, and implementation processes of various emergency plans. To ensure the comprehensiveness and authority of the data, content from government official websites, industry standardization databases, and certified authoritative news media is preferred. The crawled data is preliminarily screened and cleaned to remove redundant and irrelevant information to ensure its suitability for subsequent keyword extraction and correlation analysis.

[0060] Step 2: Preprocess the collected text data, and apply the TextRank algorithm to filter out stop words and perform word segmentation on the processed text data, so as to extract the core keywords in the text. In the specific implementation process, the emergency resource requirements involved are extracted from the historical emergency event text data, the relevant risk factors are extracted from the accident cases, and the response measures involved are identified from the emergency plan text. Through these steps, the key information extracted provides data support for the subsequent construction of the complex network.

[0061] Step 3: Use the weighted Eclat algorithm to analyze the association relationships in the cleaned text data, with a focus on mining the correlations between keywords in the text. By calculating key metrics such as support, confidence, and lift, extract rules with strong associations. These rules are used to construct a complex network model, where the extracted keywords serve as nodes in the network, and the association relationships between the keywords serve as edges in the network.

[0062] Step 4: Based on the mining results of the association rules, construct a three-layer complex network model: The first layer is the risk evolution network layer, where the nodes represent risk factors and the edges represent the evolution paths and causal relationships between risk factors; the second layer is the emergency resource network layer, where the nodes represent the emergency resources required in specific risk situations, and the edges are constructed according to the relationships between the resources; the third layer is the emergency plan network layer, where the nodes represent the response measures in the emergency plans, and the edges are constructed according to the relevance of the response processes. Further, based on the association relationships among the risk factors, emergency resource requirements, and response measures, organically connect the risk evolution network layer, the emergency resource network layer, and the emergency plan network layer to form a multi-level complex network model.

[0063] Step 5: Each node represents a specific entity, such as a risk factor, an emergency resource, or a response measure. Each node is assigned a unique serial number, and the entity represented by each serial number is stored in a document. Represent the constructed complex network in the form of node pairs and input it into the convolutional neural network training model, where text embedding or word embedding will no longer be involved. Each node serial number already has a clear meaning in the system, and this method can simplify the input and processing of data, enabling the model to focus more on learning the characteristics of the network structure.

[0064] Step 6: Input the constructed three-layer network model in the form of node pairs into the convolutional neural network model for training to learn the characteristics of the network structure. Use the stratified sampling method to divide the dataset into a training set, a validation set, and a test set to maintain a balanced distribution of edge labels. The training set accounts for 70%-80% of the total data and is used for preliminary training of the model; the validation set accounts for 10%-15% of the total data and is used for adjusting hyperparameters and optimizing the model; the test set accounts for 10%-15% of the total data and is used for evaluating the final performance of the model.

[0065] Step 7: During the training process, input positive and negative samples to optimize the model. Positive samples are the resources and response measures used in historical emergency events, and negative samples are the resources and measures not used. By distinguishing between positive and negative samples, improve the accuracy of model prediction, ensure that the output combination of emergency resources and measures is appropriate, and avoid unnecessary resource scheduling. Learn which emergency resources and response measures are the most useful and can most effectively inhibit risk evolution in a timely manner.

[0066] Step 8: After the model training is completed, the decision-making is dynamically optimized by combining multi-objective optimization algorithms. The optimization objectives include response speed, resource utilization efficiency, and cost control. Through the Pareto optimal solution set, multiple emergency response plans are generated for decision-makers to choose from.

[0067] Step 9: Evaluate the model performance using metrics such as accuracy, recall, and F1-score. According to the evaluation results, adjust the model structure, increase the data volume, or introduce more features to further improve the optimization effect of the model.

[0068] Step 10: Design the application programming interface (API) of the system to accept newly input risk data and map it to the corresponding node numbers. The system outputs the corresponding combination of emergency resources and response measures according to the trained convolutional neural network model, providing the optimal emergency response suggestions for decision-makers.

[0069] Step 11: Continuously optimize the model parameters according to the feedback of actual emergency events to improve the adaptability of the system. At the same time, with the accumulation of more historical data, expand the network nodes to ensure the uniqueness of the numbers, avoid conflicts between new nodes and existing nodes, and ensure the sustainable expansion of the system.

[0070] Through the above steps, the system provides support for the generation of emergency response strategies for the refrigerant system. The following uses a specific example to explain the content of each functional module in detail.

[0071] Data collection module: The system automatically collects historical emergency event data, accident cases, and emergency plans related to the refrigerant system from relevant databases. For example, a certain factory once suffered equipment damage and environmental pollution due to refrigerant leakage. The system obtains relevant data such as historical emergency event data, accident cases, and emergency plans related to the refrigerant system from the government emergency management system and industry databases.

[0072] Data preprocessing module: After data collection, the system cleans the text and extracts keywords. For example, the system extracts key risk factors (such as pipeline aging, temperature rise), emergency resources (such as refrigerant sealants, emergency repair teams), and response measures (such as closing refrigerant pipelines, starting backup cooling systems) from a historical emergency event of refrigerant leakage.

[0073] Association relationship mining module: Based on the preprocessed data, the system identifies the associations between key risk factors and emergency resources and response measures through association rule mining (such as support and confidence analysis). For example, the system analyzes and finds that when the pipeline is aging and accompanied by a temperature rise, starting an emergency repair team and closing the refrigerant pipeline are common response strategies.

[0074] Complex network construction module: The system constructs a three-layer complex network model based on the extracted nodes and association relationships. For an example of a refrigerant system, the first-layer nodes represent the risk factors of refrigerant leakage (such as pipeline aging, temperature increase); the second layer represents the corresponding emergency resources (such as repair teams, sealants); the third layer represents the response measures (such as closing the pipeline, repairing the pipeline). The edges represent the causal relationships or the associations of resource requirements between these nodes.

[0075] Convolutional neural network module: The constructed three-layer complex network is input into the convolutional neural network model for deep learning. By learning from historical refrigerant events, the system can identify the most effective combinations of emergency resources and measures in different risk scenarios. For example, the system learns that in the case of pipeline aging and temperature increase, starting an emergency repair team and using sealants is the best emergency response plan.

[0076] Application Programming Interface (API) module: Users can input the risk data of the current refrigerant system (such as the degree of pipeline aging, current temperature, etc.) through the Application Programming Interface (API). Based on the learning results of the convolutional neural network model, the system will output the optimal combination of emergency resources and response measures in real time. For example, it will prompt the need to mobilize a repair team and urgently close the pipeline to prevent further refrigerant leakage.

[0077] Through an example of refrigerant leakage in an industrial refrigeration system, the entire process is reinterpreted in combination with each step.

[0078] Suppose in the refrigeration system of a large factory, due to reasons such as pipeline aging and temperature increase, a crack appears in the refrigerant pipeline, increasing the risk of refrigerant leakage. To respond to this emergency in a timely manner, the enterprise hopes to use this emergency response decision optimization system to formulate a response plan and avoid the expansion of the accident.

[0079] Step 1: The system first obtains relevant data from multiple channels, including government emergency management websites, industry databases, and news media. For example, it obtains the emergency response records of historical leakage events of similar refrigerant systems from a government emergency platform and collects accident cases related to refrigerant systems from authoritative databases. In addition, it crawls the industry-standard emergency response plans, which record how to dispatch emergency resources and execute the plans during emergency events. These data are cleaned to remove redundant and irrelevant information to ensure the accuracy of subsequent processing.

[0080] Step 2: The system preprocesses the collected text data and extracts keywords using the Text Rank algorithm. In this instance, the system extracts emergency resource requirements (such as refrigerant sealants, emergency repair teams) from the text data of historical emergency events, relevant risk factors (such as pipeline aging, temperature increase) from accident cases, and response measures (such as closing refrigerant pipelines, starting backup cooling systems) from emergency plans. This information will be used for the subsequent construction of a complex network.

[0081] Step 3: The system uses the weighted Eclat algorithm to analyze the correlation relationships in the cleaned text data and discovers the associations between different keywords. For example, the system may find that "pipeline aging" and "temperature increase" are often strongly correlated with "refrigerant leakage", while "refrigerant sealants" and "emergency repair teams" are commonly used as resources and measures to address these risks. The system extracts highly correlated rules by calculating support, confidence, and lift, and these rules will form the basis of the complex network model.

[0082] Step 4: Based on the mining results of the association rules, the system constructs a three-layer complex network model: Risk evolution network layer: The nodes in this layer represent risk factors, such as "pipeline aging" and "temperature increase", and the edges between them represent the causal relationships and evolution paths of these factors; Emergency resource network layer: The nodes represent the resources that need to be mobilized under specific circumstances, such as "refrigerant sealants" and "repair teams", and the edges represent the relationships between the resources, such as the repair team using sealants to repair pipelines; Emergency plan network layer: The nodes in this layer represent the specific response measures in the emergency plan, such as "closing refrigerant pipelines" and "starting backup cooling systems", and the execution process correlations of these measures are represented by edges; The system further organically combines these three layers of networks to form a multi-level complex network model that can comprehensively describe the entire process from risk occurrence to emergency response.

[0083] Step 5: Each node represents a specific entity, such as a certain risk factor, emergency resource, or response measure. Each node is assigned a unique serial number. For example, "pipeline aging" is numbered as node 1, and "refrigerant sealants" is numbered as node 2. The correspondence between these nodes and serial numbers is stored in a document and input into the system for further analysis, without involving text embedding, enabling the model to focus more on learning the network structure.

[0084] Step 6: Input the constructed three-layer complex network model into a convolutional neural network (CNN) for training. The dataset is divided into a training set, a validation set, and a test set, keeping the label distribution of the edges balanced. For example, the system may take the combination of "pipe aging" and "refrigerant sealant" as positive samples, and some unused resource combinations as negative samples to ensure that the model can learn effective emergency response strategies.

[0085] Step 7: During the model training process, the system optimizes the model by inputting positive and negative samples. Positive samples are the resources and response measures used in actual historical emergency events. For example, the combination of "refrigerant sealant" and "closing the pipeline"; while negative samples are unused resources and measures. In this way, the system learns which resources and response measures are most effective in specific risk situations, thereby improving the prediction accuracy of the model for future emergency events.

[0086] Step 8: After the model training is completed, the system combines a multi-objective optimization algorithm for optimization. The optimization objectives include: Response speed: How to respond to emergencies quickly; Resource utilization efficiency: How to allocate and use resources most effectively; Cost control: How to reduce emergency costs while ensuring safety; The system generates multiple emergency response plans through the Pareto optimal solution set for decision-makers to choose; for example, the system may give a plan with a quick response but high cost, or a plan with a lower cost but a slightly slower response.

[0087] Step 9: The system evaluates the performance of the model through indicators such as accuracy, recall rate, and F1 score. If the output result of the model cannot accurately predict the optimal emergency resource combination, the system will adjust the model structure or increase the data volume according to the evaluation results to further optimize the model effect.

[0088] Step 10: The designed application programming interface (API) allows users to input new risk data, such as inputting parameters such as the aging degree and temperature of the current factory's pipelines. Based on the learning results of the convolutional neural network model, the system outputs the optimal combination of emergency resources and measures in real time, such as suggesting "closing the refrigerant pipeline" and "dispatching the maintenance team" to help decision-makers respond in a timely manner.

[0089] Step 11: According to the feedback of actual emergency events, the system will continuously optimize the model. For example, in an actual operation, if the refrigerant sealant is used but the effect is not ideal, the system will adjust the parameters according to this feedback. With the accumulation of more historical data, the system will also continuously expand the network nodes to ensure the sustainable scalability of the system and avoid conflicts between new nodes and existing nodes.

[0090] In summary, this example provides an emergency response decision optimization system for a refrigerant system based on artificial intelligence and complex networks and its implementation method. First, the system performs natural language processing on the text information of historical risk events and emergency response strategies, extracts key risk factors, emergency resources, and countermeasures as network nodes. According to the association rules, a network structure is constructed, where each node represents a specific entity, such as a risk factor, an emergency resource, or a countermeasure. Each node is assigned a unique serial number and recorded in the document. The relationship between nodes is represented in the form of node pairs (such as (1,5)) and input into a convolutional neural network model as structured data for training.

[0091] By learning the network structure constructed from historical data, the model can master the association between specific feature combinations (such as specific risk factors) and emergency resources and countermeasures. The input data no longer involves text or word embeddings, but is directly input based on the node serial numbers and the network structure. Since each node serial number has a clear meaning in the system, the model can simplify the data input and processing process, thus focusing on the feature learning of the network structure. This method effectively improves the decision-making efficiency of the emergency response system in a complex environment and provides a more accurate consideration basis for subsequent input matching problems.

[0092] Through this intelligent emergency response decision optimization system, when the factory faces refrigerant leakage, it can quickly obtain the optimal combination of emergency resources and countermeasures. Through multiple iterations of optimization and data accumulation, this system continuously improves its prediction ability for future emergency events, effectively reducing the impact of accidents and improving the efficiency of emergency handling.

[0093] Although the embodiments of the present invention have been shown and described, those skilled in the art should understand that various improvements, changes, substitutions, or adjustments can still be made to these embodiments without departing from the basic principles of the present invention. Therefore, the scope of the present invention should be defined by the appended claims and their equivalent technical solutions.

Claims

1. An intelligent refrigerant system emergency response decision optimization system, characterized in that: Includes the following modules: Data collection module: Automatically collect emergency plan data, historical emergency event data and risk factor data related to the refrigerant system through web crawler technology; Data preprocessing module: preprocess the text data collected by the data collection module, including removing stop words, text segmentation, and extracting keywords from the text using natural language processing (NLP) technology; the extracted keywords cover risk factors, emergency resources, and response measures; Association rule mining module: Based on the association rule mining algorithm, it conducts in-depth analysis of structured data; through the screening of support, confidence and lift, it identifies strong association rules with high confidence for the subsequent construction of multi-layer complex network models; Complex network construction module: Based on the mined association rules, a multi-layer complex network model is constructed; the multi-layer complex network model consists of three layers of nodes: the first layer of nodes represents risk factors, the second layer of nodes represents emergency resources, and the third layer of nodes represents response measures. The directed edges between nodes represent causal relationships or resource demand associations; Machine learning training module: uses convolutional neural networks, takes multi-layer complex network models as input, and maps the association between risk factors, emergency resources and response measures to a high-dimensional feature space through deep learning; the emergency response decision optimization system outputs the optimal combination of emergency resources and response measures for specific risk scenarios to form emergency response decisions; Application Programming Interface API: Provides a user interaction platform and supports real-time input of risk information; the emergency response decision optimization system automatically matches based on input information and outputs the optimal combination of emergency resources and response measures; supports dynamic data updates, adjusts decisions according to real-time situations, and continuously optimizes emergency response strategies.

2. The intelligent refrigerant system emergency response decision optimization system according to claim 1 is characterized in that: Also includes: System optimization module; The system optimization module integrates a multi-objective optimization algorithm to dynamically optimize the emergency response plan according to different emergency scenarios to meet the needs of multiple preset optimization goals; the system optimization module can dynamically adjust system parameters based on real-time feedback of emergency event data to improve the system's response capability and prediction accuracy under different risk scenarios.

3. The intelligent refrigerant system emergency response decision optimization system according to claim 1 is characterized in that: Also includes: The system expansion module automatically expands the nodes and directed edges in the complex network model as historical emergency event data and risk factor data accumulate.

4. The intelligent refrigerant system emergency response decision optimization system according to claim 1 is characterized in that: The complex network model building module builds a three-layer network model based on association rules to characterize the relationship between entities. The first layer is the risk evolution network layer, where nodes represent risk factors and directed edges represent the evolution path and causal relationship between risk factors. The second layer is the emergency resource network layer, where nodes represent emergency resources and directed edges represent the associations between resources; The third layer is the emergency plan network layer, where nodes represent response measures and directed edges represent process associations between measures; By analyzing the emergency resource demand and emergency measures corresponding to a certain risk factor in historical emergency events, the constructed risk evolution network layer, emergency resource network layer and emergency plan network layer are organically connected to form a multi-layer complex network model; Each node represents a specific entity, such as a risk factor, emergency resource or response measure; each node is assigned a unique serial number, and all entities represented by the serial numbers are stored in the document; the complex network model constructed is represented in the form of node pairs and input into the convolutional neural network for training; the input is input through the node serial number and the network structure represented by the node pair, so that the convolutional neural network focuses on learning the characteristics of the network structure; The system expansion module automatically assigns unique serial numbers to ensure that nodes remain unique and achieve seamless connection during system expansion and upgrades.

5. The intelligent refrigerant system emergency response decision optimization system according to claim 1 is characterized in that: The machine learning training module includes an input layer for receiving the constructed multi-layer complex network model. The convolutional neural network matches risk factors with emergency resources and response measures by learning the network model structure constructed in historical data.

6. The intelligent refrigerant system emergency response decision optimization system according to claim 1 is characterized in that: The network model that uses convolutional neural networks to process inputs learns the relationship patterns between nodes; the convolutional layer processes the node numbers and the corresponding relationships; Divide the input network model into training set, validation set and test set; Positive and negative samples are used in the system training process. Positive samples are the resources and measures used in historical events, and negative samples are the resources and measures that are not used, so as to improve the prediction accuracy of the system.

7. The intelligent refrigerant system emergency response decision optimization system according to claim 6 is characterized in that: By distinguishing between positive and negative samples, the resources and measures that should be used in a given accident scenario are predicted to avoid predicting unnecessary or inappropriate resources and measures.

8. The intelligent refrigerant system emergency response decision optimization system according to claim 1 is characterized in that: The system is evaluated and optimized through accuracy, recall and F1 score to ensure that the system can provide accurate emergency response decisions in new accident scenarios.

9. The intelligent refrigerant system emergency response decision optimization system according to claim 1, characterized in that: The application programming interface (API) is used to receive input accident risk information, map it to the corresponding node number, and output the optimal combination of emergency resources and response measures based on the trained convolutional neural network to provide decision makers with the best emergency response recommendations.

10. A method for optimizing emergency response decision of an intelligent refrigerant system according to any one of claims 1 to 9, characterized in that: include: Step 1: Through the data collection module, historical emergency event data, accident cases and emergency plan texts related to the refrigerant system are automatically crawled from the government emergency management platform, industry databases and authoritative news media; the collected data covers detailed records of historical events, accident descriptions and the execution process of emergency plans; the crawled data is initially screened and cleaned to remove redundant and irrelevant information; Step 2: Preprocess the cleaned text data, use the TextRank algorithm to filter stop words and segment the text data to extract core keywords; in the specific operation, extract emergency resource requirements from historical emergency event texts, extract risk factors from accident cases, and identify response measures from emergency plan texts. The extracted keyword information provides data support for the subsequent construction of complex network models; Step 3: Use the weighted Eclat algorithm to mine association rules on the processed text data and analyze the association relationship between keywords; extract strong association rules by calculating support, confidence and lift indicators, and build a complex network model based on them; keywords are used as nodes in the network model, and the association relationship between nodes is used as the edge in the network model to form a complete network structure; Step 4: Based on the mining results of association rules, a three-layer complex network model is constructed: the first layer is the risk evolution network layer, where nodes represent risk factors and directed edges represent risk evolution paths and causal relationships; the second layer is the emergency resource network layer, where nodes represent emergency resources required under specific risk scenarios and directed edges are constructed based on the correlation between resources; the third layer is the emergency plan network layer, where nodes represent specific response measures in the emergency plan and directed edges are constructed based on the logical relationship between response measures; through the association between risk factors, emergency resources and response measures, the three-layer network is integrated to form a multi-level complex network model; Step 5: Each node corresponds to a unique entity, including risk factors, emergency resources or response measures; the system assigns a unique serial number to each node and stores its specific meaning in a separate document; The constructed complex network model is input into the convolutional neural network for training in the form of node pairs; Step 6: Input the three-layer network model into the convolutional neural network in the form of node pairs, and use the stratified sampling method to divide the data set into training set, validation set and test set to ensure the balance of edge label distribution; the training set accounts for 70%-80% of the total data volume and is used for preliminary training of the system; the validation set accounts for 10%-15% and is used for hyperparameter adjustment and system optimization; the test set accounts for 10%-15% and is used to evaluate the final performance of the system; Step 7: During system training, input positive and negative samples for optimization; Positive samples are the resources and response measures actually used in historical emergency events, while negative samples are the resources and response measures that were not used. By distinguishing between positive and negative samples, the prediction accuracy of the system is improved, ensuring that the output combination of emergency resources and response measures is the most appropriate, and avoiding redundant resource scheduling. Step 8: After training is completed, the decision is optimized by combining the multi-objective optimization algorithm; the optimization objectives include response speed, resource utilization efficiency and cost control; through the Pareto optimal solution set, multiple emergency response plans are generated for decision makers to choose; Step 9: Use precision, recall, and F1 score to evaluate system performance, and adjust the system structure based on the evaluation results; Step 10: Design a system application programming interface (API) to receive newly input risk data and map it to the corresponding node number; based on the trained convolutional neural network, the system outputs the corresponding emergency resources and response measures to provide decision makers with the best emergency response recommendations; Step 11: Based on actual emergency event feedback, continuously optimize system parameters to improve system adaptability; As more historical data accumulates, the network model nodes are expanded to ensure the uniqueness of the node numbers, avoid conflicts, and ensure the sustainable expansion of the system.

Images