Emergency response grade determination method and system for hydrogenation device

Abstract

The invention discloses an emergency response grade determination method and system for a hydrogenation device, and the method comprises the steps: collecting the consequences generated when a hydrogenation device accident occurs in a historical period, the adopted emergency measures and the working data of the device, determining all the features causing the accident according to the working data, and carrying out the calculation of the emergency response grade of the hydrogenation device for each feature. Obtaining evaluation results which are made based on expert knowledge and respectively represent feature importance of the current time period and the historical time period, evaluating the reliability degree of the corresponding evaluation results according to sources, obtaining feature comprehensive importance, and generating accident association indexes; analyzing the influence of damage area types and damage ranges corresponding to different hydrogenation device accident types and emergency measures on the severity of consequences, and generating a potential consequence index; predicting working data of the target hydrogenation device, and generating an accident situation development index in combination with a fuzzy comprehensive evaluation method; and determining an emergency response level according to each index. According to the invention, the emergency response level can be accurately obtained.

Description

Technical Field

[0001] This invention belongs to the field of chemical safety technology, and in particular relates to a method and system for determining the emergency response level of a hydrogenation unit. Background Technology

[0002] Chemical production processes are complex, often involving flammable, explosive, or toxic media. Uncontrolled reactions can lead to safety accidents, endangering personal safety, property, and the environment. Addressing this issue necessitates early emergency information support across multiple stages of chemical production. However, chemical accidents are characterized by sudden occurrence, numerous uncertainties, severe consequences, and wide-ranging impacts. Therefore, early intervention in the safe operation and management of chemical plants to quickly and accurately predict the three-stage evolution of anomalies into accidents has always been a challenging and crucial research area.

[0003] With the rapid development of computer technology, the research and application of deep learning algorithms in artificial intelligence have also developed rapidly. In recent years, deep learning technology has been widely used in big data analysis, image recognition, speech recognition, video analysis, and text analysis, achieving great success. Deep learning is essentially a process of feature description. Therefore, it is feasible to conduct big data analysis based on deep learning principles and to use deep learning technology for trend prediction of key response parameters. Furthermore, this technology can be developed into a novel prediction method.

[0004] Existing technology discloses an emergency response early warning method based on power outages, including the following steps: selecting a designated area for power grid data analysis on a GIS power grid data layer and setting a threshold for the response level; collecting power grid data, meteorological data, and regional geological disaster information; establishing emergency response judgment criteria for data analysis, and determining the event type and level when at least one criterion is met; classifying the event type according to its nature and content, and classifying its specific level according to its severity; analyzing and assessing the scope of impact of the outage and the power grid equipment and facilities by combining the power grid topology and spatial geographic information; evaluating the optimal solution and developing an emergency plan. This method provides early warnings based on the level of the emergency event, and intelligent resource allocation improves the efficiency of emergency response personnel, providing strong support for emergency response. However, the emergency response judgment criteria mainly adopt qualitative and semi-quantitative methods and are only applicable to the electrical field, exhibiting strong professionalism and limitations.

[0005] Existing technology also discloses a deep learning-based method and device for predicting unstable rock deformation based on multiple time series. Its main features are: acquiring images of the unstable rock mass using a monitoring device; extracting image features using Caffe visualization tool; training an AlexNet model based on the rock mass images and feature labels to identify the distributed features of the unstable rock; constructing multiple time series of unstable rock deformation using the collected rock mass image feature data; establishing multiple data samples; fitting these samples using deep learning technology; and finally, using a screening program developed in Matlab software to optimize and compare the prediction data of multiple time series, outputting the prediction result of the time series with the smallest error. Furthermore, the corresponding device is described in detail. This method analyzes and processes deformation images of the unstable rock mass at various time stages to obtain sample data and establish a prediction model, thereby automatically and quickly predicting unstable rock deformation. It demonstrates accuracy and flexibility in unstable rock prediction, providing a basis for predicting unstable rock mass instability and for predicting and preventing rockfalls. It primarily represents the application of deep learning algorithms in the field of image recognition.

[0006] In addition, existing technologies, based on the effective approach of online monitoring and dynamic risk early warning of chemical processes to reduce accidents, combine deep learning time-series prediction models with fuzzy mathematics risk assessment models to form a novel chemical process risk early warning model. A method based on combining deep learning time-series prediction and fuzzy mathematical quantitative risk assessment is proposed. Addressing the challenges of the dynamic, temporal, and highly nonlinear nature of chemical process data, as well as the short prediction cycle, a deep learning time-series prediction model is formed by combining a convolutional neural network (CNN) with a long short-term memory (LSTM) network, achieving 108-minute advance prediction of process parameters. This method is applied to the ammonia synthesis process, predicting six risk parameters including temperature, pressure, flow rate, and hydrogen-nitrogen ratio. The prediction results show that the method has high prediction accuracy, and its linear regression correlation coefficient and root mean square error indicate very high precision. Simultaneously, the method uses triangular fuzzy numbers to assess the risk of the time-series prediction results, obtaining a time-series risk change curve, thus realizing chemical process risk early warning. This method provides a valuable exploration of using artificial intelligence and big data to achieve process control and risk early warning, offering a reference for achieving advanced early warning of chemical processes. It is mainly applied to chemical process fault prediction and diagnostic early warning.

[0007] In developing this invention, the inventors discovered that, in light of the aforementioned existing technologies, early emergency response to accidents in petrochemical enterprises' hydrogenation units has historically relied heavily on qualitative analysis methods such as fault tree and logic tree analysis, depending on historical accident data without considering the current real-time status of the hydrogenation unit. This resulted in limited accuracy in accident evolution prediction and a lack of research on early-stage trend prediction. Therefore, this invention proposes a method for determining the emergency response level of hydrogenation units that leverages the advantages of expert knowledge and deep learning algorithms. This method comprehensively considers accident correlation, potential consequences, and accident situation development, and integrates qualitative and quantitative research methods to overcome the uncertainty and diversity of key response parameters. It avoids the randomness and ambiguity inherent in the emergency response process, making it essential to determine the emergency response level of hydrogenation units comprehensively and systematically throughout the entire process. Summary of the Invention

[0008] To address the aforementioned problems, this invention provides a method for determining the emergency response level of a hydrogen refueling unit, comprising: collecting the consequences, emergency measures taken, and unit operating data of hydrogen refueling unit accidents occurring in historical periods; determining all characteristics causing the accident based on the operating data; then, for each characteristic, obtaining several first evaluation results representing the importance of the characteristics in the current period based on expert knowledge, and several second evaluation results representing the importance of the characteristics in historical periods; evaluating the reliability of each evaluation result based on its source to obtain a comprehensive importance of the characteristics, thereby generating an accident correlation index representing the comprehensive correlation between each characteristic and the target hydrogen refueling unit accident; analyzing the damage area type and damage range corresponding to different types of hydrogen refueling unit accidents, and the impact of the emergency measures on the severity of the consequences, thereby generating a potential consequence index representing the severity of the consequences of the target hydrogen refueling unit accident; obtaining the change characteristics of the operating data to predict the operating data of the target hydrogen refueling unit, and then combining this with a fuzzy comprehensive evaluation method to generate an accident situation development index representing the development trend of the target hydrogen refueling unit accident; and obtaining the current emergency response level of the target hydrogen refueling unit based on each index.

[0009] Preferably, the first evaluation result and the second evaluation result are important, unimportant, or their importance cannot be determined. The step of evaluating the reliability of each evaluation result based on its source to obtain the overall importance of the feature includes: assigning a score value to each evaluation result according to its importance; for each feature, recording the score value of each first evaluation result as a first score value; for each feature, extracting second evaluation results made by experts with experience in evaluating accident influencing factors from among the plurality of second evaluation results, recording the score value of each extracted second evaluation result as a second score value, and then determining the reliability of each extracted second evaluation result based on the professional field of the expert who made each extracted second evaluation result, and assigning a weight to the corresponding second evaluation result based on the reliability value. Based on this, combining the first score value and the second score value, a total score value is calculated to characterize the overall importance of each feature to the target hydrogenation unit accident.

[0010] Preferably, the step of generating an accident correlation index representing the comprehensive correlation between each feature and the accident of the target hydrogenation unit includes: using the working data to analyze the degree of correlation between each feature and the accident of the hydrogenation unit; further calculating the ratio between the number of related accidents and the number of unrelated accidents for each feature; using this ratio as an accident correlation correction factor to correct the accident correlation index; and based on this, combining the total score to obtain the accident correlation index.

[0011] Preferably, the total score is calculated using the following expression:

[0012]

[0013] Among them, ZS i Let f represent the total score, N represent the total number of first evaluation results, x represent the sequence number of the first evaluation result, and f represent the total score. d b1 and b2 represent coefficients, m represents the total number of first and second evaluation results, n represents the number of evaluation results that are the same as the actual evaluation result, and f represents the first rating value. j represents the second score, w represents the weight, and i represents the feature number.

[0014] Preferably, the accident correlation index is obtained using the following expression:

[0015]

[0016] Where S1 represents the accident correlation index, MS i Indicating the degree of relevance, C i This represents the accident-related correction factor, where M represents the total number of all features that cause the accident.

[0017] Preferably, the different accident types of hydrogen refueling devices include, but are not limited to: fire, explosion and poisoning, and the types of injury areas include, but are not limited to: death area, serious injury area and minor injury area. The step of analyzing the injury area type and injury range corresponding to the different accident types of hydrogen refueling devices includes: for each type of accident of hydrogen refueling device, using the environmental information of the location of the hydrogen refueling device when the accident occurred in a historical period, to obtain the injury range corresponding to the different injury area types.

[0018] Preferably, the step of analyzing the impact of the emergency measures on the severity of the consequences includes: constructing a hierarchical model based on the AHP algorithm; evaluating the impact of each emergency measure on the severity of the consequences by analyzing the first relative importance of each first evaluation indicator relative to the severity of the consequences, and the second relative importance of each second evaluation indicator relative to each of the first evaluation indicators; obtaining a first weight for evaluating the first relative importance corresponding to each first evaluation indicator, and a second weight for evaluating the second relative importance corresponding to each second evaluation indicator; thereby obtaining a weight vector for evaluating the relative importance of each second evaluation indicator relative to the severity of the consequences; and thus obtaining weight coefficients representing the impact of each evaluation indicator on the severity of the accident consequences of the target hydrogenation unit. Here, the severity of the consequences is used as the target layer element; process control, material isolation, emergency fire fighting, and area planning and fire / explosion prevention among the emergency measures are used as criterion layer elements, denoted as the first evaluation indicators; and sub-emergency measures belonging to process control, material isolation, emergency fire fighting, and area planning and fire / explosion prevention are used as measure layer elements, denoted as the second evaluation indicators.

[0019] Preferably, the step of generating a potential consequence index representing the severity of an accident at the target hydrogenation unit includes: determining the effectiveness of corresponding emergency measures based on the consequences, and assigning effective and ineffective emergency measures values ​​of 1 and 0, respectively; for each second evaluation indicator, calculating the product of the emergency measure value and the weight coefficient of the second evaluation indicator; further, for each first evaluation indicator, calculating the sum of the product results corresponding to all second evaluation indicators belonging to the first evaluation indicator, and recording it as a first result; for each first evaluation indicator, calculating the product of the sum of the first results and the weight coefficient of the first evaluation indicator, and recording it as a second result; further obtaining the sum of all second results; based on this, configuring a corresponding damage severity score value for each damage area type according to the degree of damage, and further combining it with the corresponding damage range to obtain the potential consequence index.

[0020] Preferably, the step of generating an accident situation development index representing the accident development trend of the target hydrogenation unit includes: configuring corresponding working data ranges for different probabilities of an accident occurring in the target hydrogenation unit; using a fuzzy comprehensive evaluation method to establish a correlation between the predictive working data and the predicted membership state representing the probability of an accident occurring in the target hydrogenation unit; obtaining a membership function for predicting the probability of an accident occurring in the target hydrogenation unit; and based on this, using the values ​​of the predictive working data to obtain the values ​​of the predicted membership state, thereby obtaining the accident situation development index.

[0021] Preferably, the membership function is represented by the following expression:

[0022] μ(x) = 0, x ∈ [l1, μ1]

[0023]

[0024] Where μ(x) represents the general state membership function, x represents the prediction working data, and μ 1l μ represents the membership function of the lower limit state of level I. 1h μ represents the membership function of the Level I high-limit state. 2l μ represents the membership function of the lower limit state of level II. 2h This represents the membership function for the Level II high-limit state.

[0025] Preferably, the step of acquiring the variation characteristics of the working data to predict the working data of the target hydrogenation unit includes: based on the working data, using the working data of the previous period as input information of a preset model, and using the working data of the next period as output information of the preset model, thereby constructing a working data prediction model by training the preset model; using the working data of the target hydrogenation unit in the current period as input information of the working data prediction model to predict the working data of the target hydrogenation unit, obtaining the corresponding predicted working data, and thus obtaining the predicted working data.

[0026] Preferably, in determining the reliability of each extracted second evaluation result based on the professional field to which the expert who made each extracted second evaluation result belongs, the process includes: if the professional field is a device, the reliability of the extracted second evaluation result is determined to be the highest; if the professional field is an emergency, the reliability of the extracted second evaluation result is determined to be medium; if the professional field is neither a device nor an emergency, the reliability of the extracted second evaluation result is determined to be the lowest.

[0027] Preferably, the process of assigning weights to the corresponding second evaluation results includes: assigning a weight of 2 to the second evaluation result with the highest reliability; assigning a weight of 1.5 to the second evaluation result with medium reliability; and assigning a weight of 1 to the second evaluation result with the lowest reliability.

[0028] Preferably, in the process of configuring a score value for each evaluation result according to its importance, the following steps are taken: for evaluation results that are important, unimportant, and whose importance cannot be determined, a score value of 1 point, 0.5 points, and 0 points are configured respectively.

[0029] On the other hand, the present invention also provides an emergency response level determination system for a hydrogenation unit. The system includes the following modules: an accident correlation index acquisition module, which collects the consequences, emergency measures taken, and unit operating data of accidents that occurred in hydrogenation units during historical periods. Based on the operating data, it determines all characteristics that caused the accident. Then, for each characteristic, it acquires several first evaluation results representing the importance of the characteristic in the current period, and several second evaluation results representing the importance of the characteristic in historical periods, based on expert knowledge. Based on the source of each evaluation result, it evaluates the reliability of each evaluation result to obtain the comprehensive importance of the characteristic, thereby generating a value representing the overall importance of each characteristic. The system comprises: an accident correlation index representing the comprehensive correlation between characteristics and the target hydrogenation unit accident; a potential consequence index acquisition module, which analyzes the damage area type and damage range corresponding to different hydrogenation unit accident types, as well as the impact of emergency measures on the severity of the consequences, to generate a potential consequence index representing the severity of the target hydrogenation unit accident consequences; an accident situation development index acquisition module, which acquires the change characteristics of the operating data to predict the operating data of the target hydrogenation unit, and then combines the fuzzy comprehensive evaluation method to generate an accident situation development index representing the development trend of the target hydrogenation unit accident; and an emergency response level determination module, which obtains the current emergency response level of the target hydrogenation unit based on the indices.

[0030] Compared with the prior art, one or more embodiments of the above solutions may have the following advantages or beneficial effects:

[0031] This invention proposes a method and system for determining the emergency response level of a hydrogen refueling unit. The method first collects data related to accidents occurring in hydrogen refueling units during historical periods (consequences, emergency measures taken, and unit operating data). Next, based on the operating data, all characteristics causing the accident are identified. Then, for each characteristic, evaluation results based on expert knowledge are obtained, representing the importance of the characteristic in the current and historical periods respectively. The reliability of each evaluation result is evaluated based on its source to obtain a comprehensive characteristic importance, generating an accident correlation index. Then, the damage area type and damage range corresponding to different types of hydrogen refueling unit accidents are analyzed, as well as the impact of emergency measures on the severity of consequences, to generate a potential consequence index representing the severity of the accident consequences of the target hydrogen refueling unit. Next, the changing characteristics of the operating data are acquired to predict the operating data of the target hydrogen refueling unit. Then, combined with a fuzzy comprehensive evaluation method, an accident situation development index representing the development trend of the accident in the target hydrogen refueling unit is generated. Finally, the emergency response level of the current target hydrogen refueling unit is obtained based on each index. This invention effectively combines the advantages of expert knowledge and deep learning algorithms, comprehensively considering the correlation between features and accidents, the potential consequences of accidents, and the development of accident situations. It integrates qualitative and quantitative research methods to accurately obtain the emergency response level. Furthermore, this invention overcomes the uncertainty and diversity of the characteristics causing accidents in hydrogenation units, effectively avoiding randomness and ambiguity in the emergency response process.

[0032] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the description, claims, and drawings. Attached Figure Description

[0033] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0034] Figure 1 This is a step diagram illustrating the method for determining the emergency response level of a hydrogenation unit according to an embodiment of this application.

[0035] Figure 2 This is a block diagram of an emergency response level determination system for a hydrogenation unit, as described in an embodiment of this application. Detailed Implementation

[0036] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings and examples, so that the process of how the present invention uses technical means to solve technical problems and achieve technical effects can be fully understood and implemented accordingly. It should be noted that, as long as there is no conflict, the various embodiments and features in the various embodiments of the present invention can be combined with each other, and the resulting technical solutions are all within the protection scope of the present invention.

[0037] Furthermore, the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0038] Chemical production processes are complex, often involving flammable, explosive, or toxic media. Uncontrolled reactions can lead to safety accidents, endangering personal safety, property, and the environment. Addressing this issue necessitates early emergency information support across multiple stages of chemical production. However, chemical accidents are characterized by sudden occurrence, numerous uncertainties, severe consequences, and wide-ranging impacts. Therefore, early intervention in the safe operation and management of chemical plants to quickly and accurately predict the three-stage evolution of anomalies into accidents has always been a challenging and crucial research area.

[0039] With the rapid development of computer technology, the research and application of deep learning algorithms in artificial intelligence have also developed rapidly. In recent years, deep learning technology has been widely used in big data analysis, image recognition, speech recognition, video analysis, and text analysis, achieving great success. Deep learning is essentially a process of feature description. Therefore, it is feasible to conduct big data analysis based on deep learning principles and to use deep learning technology for trend prediction of key response parameters. Furthermore, this technology can be developed into a novel prediction method.

[0040] Existing technology discloses an emergency response early warning method based on power outages, including the following steps: selecting a designated area for power grid data analysis on a GIS power grid data layer and setting a threshold for the response level; collecting power grid data, meteorological data, and regional geological disaster information; establishing emergency response judgment criteria for data analysis, and determining the event type and level when at least one criterion is met; classifying the event type according to its nature and content, and classifying its specific level according to its severity; analyzing and assessing the scope of impact of the outage and the power grid equipment and facilities by combining the power grid topology and spatial geographic information; evaluating the optimal solution and developing an emergency plan. This method provides early warnings based on the level of the emergency event, and intelligent resource allocation improves the efficiency of emergency response personnel, providing strong support for emergency response. However, the emergency response judgment criteria mainly adopt qualitative and semi-quantitative methods and are only applicable to the electrical field, exhibiting strong professionalism and limitations.

[0041] Existing technology also discloses a deep learning-based method and device for predicting unstable rock deformation based on multiple time series. Its main features are: acquiring images of the unstable rock mass using a monitoring device; extracting image features using Caffe visualization tool; training an AlexNet model based on the rock mass images and feature labels to identify the distributed features of the unstable rock; constructing multiple time series of unstable rock deformation using the collected rock mass image feature data; establishing multiple data samples; fitting these samples using deep learning technology; and finally, using a screening program developed in Matlab software to optimize and compare the prediction data of multiple time series, outputting the prediction result of the time series with the smallest error. Furthermore, the corresponding device is described in detail. This method analyzes and processes deformation images of the unstable rock mass at various time stages to obtain sample data and establish a prediction model, thereby automatically and quickly predicting unstable rock deformation. It demonstrates accuracy and flexibility in unstable rock prediction, providing a basis for predicting unstable rock mass instability and for predicting and preventing rockfalls. It primarily represents the application of deep learning algorithms in the field of image recognition.

[0042] In addition, existing technologies, based on the effective approach of online monitoring and dynamic risk early warning of chemical processes to reduce accidents, combine deep learning time-series prediction models with fuzzy mathematics risk assessment models to form a novel chemical process risk early warning model. A method based on combining deep learning time-series prediction and fuzzy mathematical quantitative risk assessment is proposed. Addressing the challenges of the dynamic, temporal, and highly nonlinear nature of chemical process data, as well as the short prediction cycle, a deep learning time-series prediction model is formed by combining a convolutional neural network (CNN) with a long short-term memory (LSTM) network, achieving 108-minute advance prediction of process parameters. This method is applied to the ammonia synthesis process, predicting six risk parameters including temperature, pressure, flow rate, and hydrogen-nitrogen ratio. The prediction results show that the method has high prediction accuracy, and its linear regression correlation coefficient and root mean square error indicate very high precision. Simultaneously, the method uses triangular fuzzy numbers to assess the risk of the time-series prediction results, obtaining a time-series risk change curve, thus realizing chemical process risk early warning. This method provides a valuable exploration of using artificial intelligence and big data to achieve process control and risk early warning, offering a reference for achieving advanced early warning of chemical processes. It is mainly applied to chemical process fault prediction and diagnostic early warning.

[0043] In developing this invention, the inventors discovered that, in light of the aforementioned existing technologies, early emergency response to accidents in petrochemical enterprises' hydrogenation units has historically relied heavily on qualitative analysis methods such as fault tree and logic tree analysis, depending on historical accident data without considering the current real-time status of the hydrogenation unit. This resulted in limited accuracy in accident evolution prediction and a lack of research on early-stage trend prediction. Therefore, this invention proposes a method for determining the emergency response level of hydrogenation units that leverages the advantages of expert knowledge and deep learning algorithms. This method comprehensively considers accident correlation, potential consequences, and accident situation development, and integrates qualitative and quantitative research methods to overcome the uncertainty and diversity of key response parameters. It avoids the randomness and ambiguity inherent in the emergency response process, making it essential to determine the emergency response level of hydrogenation units comprehensively and systematically throughout the entire process.

[0044] Therefore, to address the aforementioned problems, this invention proposes a method and system for determining the emergency response level of a hydrogen refueling unit. The method first collects data related to accidents occurring in hydrogen refueling units during historical periods (consequences, emergency measures taken, and unit operating data). Next, based on the operating data, all characteristics causing the accident are identified. Then, for each characteristic, evaluation results based on expert knowledge are obtained, representing the importance of the characteristics in the current and historical periods respectively. The reliability of each evaluation result is evaluated based on its source to obtain a comprehensive characteristic importance, generating an accident correlation index. Then, the types and extent of damage corresponding to different types of hydrogen refueling unit accidents, as well as the impact of emergency measures on the severity of consequences, are analyzed to generate a potential consequence index representing the severity of the accident consequences of the target hydrogen refueling unit. Next, the changing characteristics of the operating data are acquired to predict the operating data of the target hydrogen refueling unit. Then, combined with a fuzzy comprehensive evaluation method, an accident situation development index representing the development trend of the accident in the target hydrogen refueling unit is generated. Finally, the emergency response level of the current target hydrogen refueling unit is obtained based on each index. This invention effectively combines the advantages of expert knowledge and deep learning algorithms, comprehensively considering the correlation between features and accidents, the potential consequences of accidents, and the development of accident situations. It integrates qualitative and quantitative research methods to accurately obtain the emergency response level. Furthermore, this invention overcomes the uncertainty and diversity of the characteristics causing accidents in hydrogenation units, effectively avoiding randomness and ambiguity in the emergency response process.

[0045] Example 1

[0046] Figure 1 This is a step diagram illustrating the method for determining the emergency response level of a hydrogenation unit according to an embodiment of this application. See below for reference. Figure 1 The various steps of the present invention will be explained below.

[0047] like Figure 1 As shown, in step S110, the consequences, emergency measures and operating data of hydrogen refueling unit accidents during historical periods are collected. Based on the operating data, all characteristics that caused the accident are determined. Then, for each characteristic, several first evaluation results representing the importance of the characteristics in the current period and several second evaluation results representing the importance of the characteristics in historical periods are obtained based on expert knowledge. The reliability of each evaluation result is evaluated according to the source of each evaluation result to obtain the comprehensive importance of the characteristics, so as to generate an accident correlation index representing the comprehensive correlation between each characteristic and the target hydrogen refueling unit accident.

[0048] Specifically, this embodiment first collects historical data related to hydrogen refueling unit accidents during historical periods, including the consequences, emergency measures taken, and unit operating data. Then, by analyzing the changes in each type of operating data during the accidents in historical periods, all characteristics causing the accidents are obtained. In practical applications, when data evaluation is applied to machine learning, the importance of features depends on the expert's domain knowledge. Therefore, this embodiment quantifies expert knowledge through weighted scoring and integrates the quantification results into the feature acquisition process. Statistical analysis methods and information theory are used to analyze the relationship between features and target attributes, thereby eliminating the risk of key features being deleted by removing the most redundant or irrelevant features associated with the target attributes. For each feature, several first evaluation results representing the importance of the feature in the current period and several second evaluation results representing the importance of the feature in historical periods are obtained from different experts. Since there are usually differences in experience and professional fields among the experts making each evaluation result, the reliability of each evaluation result also varies. Accordingly, this embodiment continues to evaluate the reliability of each evaluation result based on the experience and professional fields among the experts making each evaluation result (i.e., the source of each evaluation result). Subsequently, the overall importance of the features is obtained by assigning values ​​to the importance and reliability represented by each evaluation result. Finally, an accident correlation index is generated, representing the overall correlation between each feature and an accident at the target hydrogenation unit.

[0049] In the step of evaluating the reliability of each evaluation result based on its source and obtaining the overall importance of the feature, firstly, a score value is assigned to each evaluation result according to its importance; then, for each feature, the score value of each first evaluation result is recorded as the first score value; finally, for each feature, several second evaluation results made by experts with experience in evaluating accident influencing factors are extracted, and the score value of each extracted second evaluation result is recorded as the second score value. Then, based on the professional field of the expert who made each extracted second evaluation result, the reliability of each extracted second evaluation result is determined, and a weight is assigned to the corresponding second evaluation result according to the reliability. Based on this, the first score value and the second score value are combined to calculate the total score value to characterize the overall importance of each feature to the accident of the target hydrogenation unit.

[0050] In one specific embodiment of this application, the first evaluation result and the second evaluation result are considered important, unimportant, or their importance cannot be determined. First, for each feature, a score is assigned to each first and second evaluation result, considering only the importance represented by the evaluation result itself, without considering the source. Next, for each feature, based on the aforementioned score assigned to each evaluation result, a score representing the current importance of that feature to the accident is obtained, denoted as the first score. Then, for each feature, several second evaluation results are extracted from those made by experts with experience in evaluating accident influencing factors. Based on this, and according to the aforementioned score assigned to each evaluation result, the score of each extracted second evaluation result is used as a score representing the historical importance of that feature to the accident, denoted as the second score. Further, based on the professional field of the expert who made each extracted second evaluation result, the reliability of each extracted second evaluation result is determined. Based on the reliability, a score value that considers only the source, without considering the importance represented by the evaluation result itself, is assigned as a weight for the corresponding second evaluation result, thereby completing the weight assignment. Finally, using the weighted values, the first score, and the second score, the total score representing the overall importance of each feature to the accident of the target hydrogenation unit is calculated.

[0051] In one specific embodiment of this application, during the process of configuring a score value for each evaluation result according to its importance, without considering the source but only the importance represented by the evaluation result itself, a score value of 1 point, 0.5 points, and 0 points are respectively configured for the evaluation result as important, unimportant, and uncertain whether it is important.

[0052] In one specific embodiment of this application, in determining the reliability of each extracted second evaluation result based on the professional field of the expert who made each extracted second evaluation result, if the professional field of the expert who made the second evaluation result is device, the reliability of the extracted second evaluation result is determined to be the highest; if the professional field of the expert who made the second evaluation result is emergency response, the reliability of the extracted second evaluation result is determined to be medium; if the professional field of the expert who made the second evaluation result is neither device nor emergency response, the reliability of the extracted second evaluation result is determined to be the lowest.

[0053] In one specific embodiment of this application, during the process of assigning weights to the corresponding second evaluation results, without considering the importance represented by the evaluation results themselves but only considering the source, according to the principle that the larger the weight value, the more important the second evaluation result, the weight of the second evaluation result with the highest reliability is assigned to 2; the weight of the second evaluation result with medium reliability is assigned to 1.5; and the weight of the second evaluation result with the lowest reliability is assigned to 1.

[0054] In this embodiment, without considering the evaluation result itself but only the data source, the weight is assigned as follows: the larger the weight, the more important the source of the second evaluation result. The weights are 2 for the highest reliability, 1.5 for the medium reliability, and 1 for the lowest reliability. That is, if the second evaluation result has the highest reliability, the weight of the second evaluation result is assigned to 2; if the second evaluation result has the medium reliability, the weight of the second evaluation result is assigned to 1.5; and if the second evaluation result has the lowest reliability, the weight of the second evaluation result is assigned to 1.

[0055] In this embodiment of the application, the total score is calculated using the following expression:

[0056]

[0057] Among them, ZS i Let f represent the total score, N represent the total number of first evaluation results, x represent the sequence number of the first evaluation result, and f represent the total score. d b1 and b2 represent coefficients, m represents the total number of first and second evaluation results, n represents the number of evaluation results that are the same as the actual evaluation result, and f represents the first rating value. j represents the second score, w represents the weight, and i represents the feature number.

[0058] In the embodiments of this application, The score representing the comprehensive importance of historical period characteristics, as determined by all experts with experience in evaluating accident influencing factors, is a weighted average of the second score values ​​corresponding to several second evaluation results for the respective characteristics. In other words, this embodiment essentially assigns a first weight to each first evaluation result. And assigning a second weight to historical evaluation results that represent the overall importance of the aforementioned single feature to the accident during a historical period. The calculation of the total score is implemented, where the first weight (less than 1 / 2) is always less than the second weight (greater than 1 / 2), and the total score should range from 0 to 1.

[0059] In one specific embodiment of this application, b1 and b2 are constants, with b1 less than or equal to b2. The specific values ​​are determined based on the richness of experimental data and expert experience, and are used to distinguish the credibility of actual evaluation results from historical evaluation results. Since the first weight is obtained based on a scoring method that only considers the importance represented by the evaluation result itself without considering the source, while the second weight is obtained based on a scoring method that considers both the importance represented by the evaluation result itself and the source, the first evaluation result has lower credibility than the second evaluation result. Therefore, the preferred value for b1 is 9, and the preferred value for b2 is 8.

[0060] In the step of generating an accident correlation index that represents the comprehensive correlation between each feature and the accident of the target hydrogenation unit, the correlation between each feature and the accident of the hydrogenation unit is analyzed using working data. Furthermore, the ratio between the number of relevant accidents and the number of unrelated accidents is calculated for each feature, and this ratio is used as an accident correlation correction factor to correct the accident correlation index. Based on this, combined with the total score, the accident correlation index is obtained.

[0061] Specifically, this embodiment utilizes collected working data and employs methods such as curve fitting to analyze the correlation between each feature and accidents in the hydrogenation unit, and calculates the correlation coefficient to characterize the corresponding correlation. Then, the number of accidents related to and unrelated to each feature is statistically analyzed, i.e., the number of related accidents and the number of unrelated accidents. Furthermore, the ratio between the number of related accidents and the number of unrelated accidents is calculated for each feature. This ratio is then used as an accident correlation correction factor to adjust the accident correlation index. Finally, the accident correlation index is calculated using the correlation degree, the total score, and the accident correlation correction factor.

[0062] In this embodiment of the application, the accident correlation index is obtained using the following expression:

[0063]

[0064] Where S1 represents the accident correlation index, MS i Indicating the degree of relevance, C i This represents the accident-related correction factor, where M represents the total number of all features that cause the accident.

[0065] In one specific embodiment of this application, the degree of relevance is characterized using feature importance scores obtained based on a machine learning model.

[0066] Furthermore, in step S120, the types and extent of damage corresponding to different types of hydrogen refueling unit accidents are analyzed, as well as the impact of emergency measures on the severity of the consequences, to generate a potential consequence index representing the severity of the consequences of the target hydrogen refueling unit accident. In this embodiment, hydrogen refueling unit accidents occurring in historical periods are used as research samples. A logic tree is used for accident analysis and organization. Based on the accident occurrence mechanism, the types of hydrogen refueling unit accidents are generally classified. For each type of hydrogen refueling unit accident, the types and extent of damage corresponding to different accident types are analyzed, as well as the impact of emergency measures on the severity of the consequences, to obtain the comprehensive consequences that the target hydrogen refueling unit accident may produce, thereby generating a potential consequence index representing the severity of the consequences of the target hydrogen refueling unit accident.

[0067] In practical applications, hydrogen refueling unit accidents are mainly classified into three categories: fire, explosion, and poisoning. The consequences are calculated based on corresponding gas leakage diffusion models, fire thermal radiation analysis models, and explosion shock wave analysis models. Specifically, the concentration of leaked toxic substances is taken from the gas leakage diffusion model, the heat flux value from the fire thermal radiation analysis model, and the calculated values ​​for the fatal radius, serious injury radius, and minor injury radius of an explosion accident are taken from the explosion shock wave analysis model. Therefore, in a specific embodiment of this application, different accident types for hydrogen refueling units include, but are not limited to, fire, explosion, and poisoning. The types of injury areas include, but are not limited to, fatal areas, seriously injured areas, and slightly injured areas.

[0068] In the step of analyzing the injury zone type and injury range corresponding to different types of hydrogen refueling unit accidents, for each type of hydrogen refueling unit accident, the injury range is obtained by utilizing environmental information about the location of the hydrogen refueling unit at the time of the accident in historical periods. Specifically, in this embodiment, for each type of hydrogen refueling unit accident, the injury radius is calculated from the center point of the unit to the end of the death zone, from the end of the death zone to the end of the seriously injured zone, and from the end of the seriously injured zone to the end of the injury zone, thereby obtaining the injury range. The probability of death within the death zone is 50%, the probability of death within the seriously injured zone is 30%, and the probability of death within the minor injury zone is 10%.

[0069] In one specific embodiment of this application, for a fire, the formation of the damage zone is mainly related to thermal radiation, and the degree of injury within the fire damage zone can be quantitatively represented by a probabilistic equation. The probabilistic equation uses the PROBIT model to give the relationship between the magnitude of the heat flux and the lethal probability, transforming the relationship curve into an equivalent straight line. Using the following PROBIT model, the corresponding damage radius can be obtained by determining the corresponding heat flux based on the preset lethal probability:

[0070] P r热 = -36.38 + 2.56ln(q) 4 / 3 ×t)(4)

[0071] Among them, P r热 Let t represent the probability of death from fire, q represent the exposure time, and q represent the heat flux.

[0072] In one specific embodiment of this application, for an explosion, the formation of the damage zone is mainly related to the destruction of the shock wave. When obtaining the damage radius of the death zone, the overpressure-impulse criterion is adopted; when obtaining the damage radius of the severely injured zone and the slightly injured zone, the overpressure criterion is adopted.

[0073] In obtaining the damage radius of the explosion's death zone, by establishing the relationship between the damage radius and the explosion amount, the corresponding damage radius can be determined based on the preset explosion amount. The relationship between the damage radius and the explosion amount is expressed by the following expression:

[0074]

[0075] Where R1 represents the damage radius of the explosion's death zone, W TNT This indicates the TNT equivalent mass of the material.

[0076] In obtaining the damage radius of the severely or lightly injured area of ​​an explosion, by establishing the relationship between the damage radius and the shock wave overpressure, the corresponding damage radius can be determined based on the preset shock wave overpressure. The relationship between the damage radius and the shock wave overpressure is expressed by the following expression:

[0077] ΔP = 0.137Z -3 +0.119Z -2 +0.267Z -1 -0.019(6)

[0078]

[0079] E=W TNT Q TNT (8)

[0080] Where ΔP represents the shock wave overpressure, Z represents the intermediate variable, R2 represents the damage radius of the severely or lightly injured area of ​​the explosion, p0 represents the ambient pressure, E represents the total explosion energy, and Q... TNT This refers to the energy released when TNT explodes.

[0081] In one specific embodiment of this application, the average energy released during a TNT explosion is 4500 KJ / Kg.

[0082] In one specific embodiment of this application, for poisoning, the probability of death under toxic exposure can be determined by using the following expression to obtain the corresponding lesion radius based on the concentration, according to a preset probability of death:

[0083] P r毒 =a+bln(C c t)(9)

[0084] Among them, P r毒 The denot represents the probability of death from poisoning, a, b, and c represent constants used to describe the toxicity of the substance, C represents the concentration, and t represents the exposure time.

[0085] In the step of assessing the impact of emergency measures on the severity of consequences, a hierarchical model is constructed based on the AHP algorithm. By analyzing the first relative importance of each first evaluation indicator relative to the severity of consequences, and the second relative importance of each second evaluation indicator relative to each of the first evaluation indicators, the impact of each emergency measure on the severity of consequences is evaluated. This yields a first weight for evaluating the first relative importance of each first evaluation indicator and a second weight for evaluating the second relative importance of each second evaluation indicator. This results in a weight vector for evaluating the relative importance of each second evaluation indicator relative to the severity of consequences, thus obtaining weight coefficients representing the impact of each evaluation indicator on the severity of the accident consequences of the target hydrogenation unit. Here, the severity of consequences is used as the target layer element; process control, material isolation, emergency fire fighting, and regional planning and fire / explosion prevention among the emergency measures are used as the criterion layer elements, denoted as the first evaluation indicators; and sub-emergency measures belonging to process control, material isolation, emergency fire fighting, and regional planning and fire / explosion prevention are used as the measure layer elements, denoted as the second evaluation indicators.

[0086] Specifically, this embodiment employs the AHP algorithm. By comparing each pair of first evaluation indicators, it determines the importance of one first evaluation indicator relative to the other within the target layer elements, thus determining the importance of each compared first evaluation indicator and obtaining the first relative importance of each first evaluation indicator. Then, it continues by comparing each pair of second evaluation indicators to determine the importance of one second evaluation indicator relative to the other within each of the first evaluation indicators, thus obtaining the second relative importance of each second evaluation indicator. Based on the first and second relative importance of each first and second evaluation indicator, the relative importance of each second evaluation indicator relative to the target layer elements is obtained. Accordingly, this embodiment hierarchically processes the first and second relative importance to be analyzed, constructing a hierarchical structure model. It uses target layer and criterion layer elements to analyze the first relative importance, and uses criterion layer and measure layer elements to analyze the second relative importance, ultimately obtaining the weighted assignment results for evaluating the relative importance of each second evaluation indicator to the severity of the consequences. The target layer elements include the severity of consequences, the criteria layer elements include each primary evaluation indicator, and the measures layer elements include each secondary evaluation indicator.

[0087] In obtaining the weight assignment results, a first judgment matrix for obtaining the first weight is constructed using the rating measure of the first relative importance of each first evaluation indicator. After obtaining the first judgment matrix, its approximate eigenvector is obtained. Then, based on the approximate eigenvector of the first judgment matrix, a first weight vector for evaluating the first relative importance of each first evaluation indicator is obtained, thus obtaining the corresponding first weight. A second judgment matrix for obtaining the second weight is constructed using the rating measure of the second relative importance of each second evaluation indicator. After obtaining the second judgment matrix, its approximate eigenvector is obtained. Then, the approximate eigenvector of the second judgment matrix is ​​used to obtain a second weight vector for evaluating the second relative importance of each second evaluation indicator, thus obtaining the corresponding second weight. The first weight is coupled with each second weight under the corresponding first evaluation indicator. First, the product of each second weight and its corresponding first weight is calculated. Then, the product calculation result for each second evaluation indicator is used as the weight vector for evaluating the relative importance of each second evaluation indicator relative to the severity of the consequences, thus obtaining the corresponding weight assignment results. At this point, the weight coefficient of each first evaluation indicator will be obtained according to the first weight, and the weight coefficient of each second evaluation indicator will be obtained according to the weight assignment result. Thus, the weight coefficient representing the impact of each evaluation indicator on the severity of the accident consequences of the target hydrogenation unit is obtained.

[0088] In one specific embodiment of this application, process control, material isolation, emergency fire protection, and zoning planning and fire / explosion prevention are used as criteria layer elements and are denoted as the first evaluation index. Emergency power supplies, cooling devices, explosion suppression devices, emergency shut-off devices, computer-controlled process control, inert gas protection, process operation, chemically reactive substance inspection, and other process hazard analyses subordinate to process control; remote control (e.g., remote control valves), unloading / venting devices, emission systems, and interlocking devices subordinate to material isolation; fire power distribution systems, fire water supply systems, automatic fire alarm systems, fire cooling water systems, water mist extinguishing systems, foam extinguishing systems, fire hydrant systems, fire monitors, hand-held fire extinguishers, water spray guns, and water curtains subordinate to emergency fire protection; and device location layout, internal layout of devices, safe evacuation and refuge, fire-resistant construction, fire resistance rating, fire compartmentation, explosion-proof facilities, and electrical fire protection facilities subordinate to zoning planning and fire / explosion prevention are used as measures layer elements and are denoted as the second evaluation index.

[0089] In the step of generating a potential consequence index representing the severity of an accident at a target hydrogenation unit, the effectiveness of corresponding emergency measures is first determined based on the consequences, with effective and ineffective emergency measures assigned values ​​of 1 and 0, respectively. Then, for each second evaluation indicator, the product of the assigned emergency measure value and the weight coefficient of the second evaluation indicator is calculated. Further, for each first evaluation indicator, the sum of the products of all second evaluation indicators belonging to the first evaluation indicator is calculated and recorded as the first result. Finally, for each first evaluation indicator, the product of the sum of the first results and the weight coefficient of the first evaluation indicator is calculated and recorded as the second result. The sum of all second results is then obtained. Based on this, a corresponding damage severity score is assigned to each damage area type according to the degree of damage, and further combined with the corresponding damage range to obtain the potential consequence index.

[0090] Specifically, this embodiment quantifies the actual impact of each emergency measure on the consequences by assigning values ​​to the emergency measures, that is, the actual effectiveness of each evaluation indicator. Effective emergency measures are assigned values ​​of 1 and ineffective emergency measures, respectively. Next, the product of the assigned emergency measure value and the weight coefficient under each second evaluation indicator is calculated. Furthermore, the products of all second evaluation indicators under each first evaluation indicator are summed to obtain a summation result appropriate to each first evaluation indicator, denoted as the first result. Finally, the product of the first result and the weight coefficient under each first evaluation indicator is calculated, denoted as the second result, and all second results are summed. At this point, the degree of injury for each injury area type is quantified by assigning values, where injury degree scores of 0.5, 0.15, and 0.05 are assigned to death areas, severely injured areas, and slightly injured areas, respectively. Based on this, the product between the damage radius of the damage area type corresponding to each type of hydrogen refueling unit accident and the corresponding damage severity score is calculated, and all product results for all hydrogen refueling unit accident types are summed. Then, by calculating the product of the summation of all second results and the summation of all product results for all hydrogen refueling unit accident types regarding the damage radius, the potential consequence index is obtained.

[0091] Further, in step S130, the changing characteristics of the working data are acquired to predict the working data of the target hydrogenation unit. Then, using the fuzzy comprehensive evaluation method, an accident situation development index representing the accident development trend of the target hydrogenation unit is generated. Specifically, this embodiment first analyzes the changing trend of the working data to obtain the corresponding changing characteristics. Then, based on the changing characteristics, the changing trend of the working data of the target hydrogenation unit is determined, thereby obtaining the predicted working data of the target hydrogenation unit. Next, the fuzzy comprehensive evaluation method is introduced to establish the correlation between the predicted working data and the accident situation development index, thereby achieving the purpose of generating an accident situation development index representing the accident development trend of the target hydrogenation unit.

[0092] In the step of acquiring the changing characteristics of working data to predict the working data of the target hydrogenation unit, firstly, based on the working data, the working data of the previous period is used as the input information of the preset model, and the working data of the next period is used as the output information of the preset model, thereby constructing a working data prediction model through training the preset model; then, the working data of the target hydrogenation unit in the current period is used as the input information of the working data prediction model to predict the working data of the target hydrogenation unit, thereby obtaining the corresponding predicted working data, and finally obtaining the predicted working data.

[0093] Specifically, this embodiment constructs a preset model based on machine learning technology. Then, the operating data from the previous time period is used as input to the preset model, and the operating data from the next time period is used as output to train the model. After training, an operating data prediction model is obtained. After obtaining the operating data prediction model, the operating data of the target hydrogenation unit in the current time period is used as input to predict the operating data for that period. In other words, the output data of the operating data prediction model at this time is the predicted operating data. By analyzing the trend of the predicted operating data over time, the corresponding predicted operating data can be obtained. Therefore, this invention, by analyzing the trend of the characteristics causing accidents in hydrogenation units over time to obtain sample data for training the preset model, establishes an operating data prediction model. This model can automatically and quickly predict the changing characteristics of accidents causing accidents in target hydrogenation units, laying a solid data foundation for effectively quantifying the development of accident situations.

[0094] In one specific embodiment of this application, before using the working data from the previous period as input to the preset model, the working data used to train the preset model is divided into a training set and a test set according to a specified ratio (e.g., 75% of the working data used to train the preset model forms the training set, and 25% forms the test set). The data in the training set is used to train the preset model; the data in the test set is used to evaluate the prediction performance of the final working data prediction model. If the evaluation result does not match the expectation, retraining is performed to ensure the prediction accuracy and reliability of the working data prediction model used to predict the working data of the target hydrogenation unit.

[0095] In the step of generating the accident situation development index representing the accident development trend of the target hydrogenation unit, firstly, corresponding working data ranges are configured for different possible degrees of accident occurrence of the target hydrogenation unit. Then, the fuzzy comprehensive evaluation method is used to establish the correlation between the predictive working data and the predicted membership state representing the probability of the target hydrogenation unit causing an accident, thereby obtaining the membership function used to predict the probability of the target hydrogenation unit causing an accident. Based on this, the numerical values ​​of the predicted membership state are obtained using the numerical values ​​of the predictive working data, and thus the accident situation development index is obtained.

[0096] Specifically, this embodiment, based on the alarm thresholds of the enterprise's DCS system, uses expert knowledge and other methods to configure working data ranges adapted to different probabilities of accidents occurring at the target hydrogen refueling unit. Then, using fuzzy comprehensive evaluation, the predicted working data and the predicted membership state representing the probability of an accident at the target hydrogen refueling unit are used as independent and dependent variables, respectively. Combined with the configured working data ranges, a membership function for predicting the probability of an accident at the target hydrogen refueling unit is constructed. Finally, based on the working data range to which the predicted working data value belongs, the predicted working data value is substituted into the membership function to calculate the predicted membership state value. Then, based on the predicted membership state value, the accident situation development index is obtained using the correspondence between the accident situation development index and the predicted membership state value, as shown in Table 1. In addition, this embodiment also provides an alarm based on the accident situation development index according to actual needs.

[0097] In this embodiment, the membership functions (general state membership function, Level I lower limit state membership function, Level I higher limit state membership function, Level II lower limit state membership function, and Level II higher limit state membership function) are represented by the following expressions:

[0098] μ(x)=0,x∈[l1,μ1](10)

[0099]

[0100]

[0101] Where μ(x) represents the general state membership function, x represents the prediction working data, and μ 1l μ represents the membership function of the lower limit state of level I. 1h μ represents the membership function of the Level I high-limit state. 2l μ represents the membership function of the lower limit state of level II. 2h This represents the membership function for the Level II high-limit state.

[0102] Table 1. Accident Situation Development Index Classification Table

[0103]

[0104]

[0105] Further, in step S140, the emergency response level of the current target hydrogenation unit is obtained based on each index. Specifically, in this embodiment, after obtaining the accident correlation index, potential consequence index, and accident situation development index, the emergency response level is determined based on preset emergency response level determination conditions. For example, based on the preset emergency response level and the corresponding threshold range combination for the accident correlation index, potential consequence index, and accident situation development index, the emergency response level corresponding to the threshold range combination that simultaneously satisfies the current accident correlation index, potential consequence index, and accident situation development index is taken as the emergency response level of the current target hydrogenation unit.

[0106] In one specific embodiment of this application, the preset emergency response level is a three-level emergency response, including: a Level 1 abnormal emergency response indicating that the hydrogenation unit has abnormal accident characteristics but the accident has not yet occurred; a Level 2 initial emergency response indicating that the accident has occurred but no emergency measures have been taken; and a Level 3 accident emergency response indicating that emergency measures have been taken but the accident situation has further expanded.

[0107] Example 2

[0108] Based on the emergency response level determination method for hydrogenation units described in Embodiment 1 above, this invention also provides an emergency response level determination system for hydrogenation units (hereinafter referred to as the "emergency response level determination system"). Figure 2 This is a block diagram of an emergency response level determination system for a hydrogenation unit, as described in an embodiment of this application.

[0109] like Figure 2As shown, the emergency response level determination system in this embodiment of the invention includes: an accident correlation index acquisition module 21, a potential consequence index acquisition module 22, an accident situation development index acquisition module 23, and an emergency response level determination module 24. Specifically, the accident correlation index acquisition module 21 is implemented according to the method described in step S110 above, configured to collect the consequences, emergency measures taken, and operating data of the hydrogenation unit when accidents occurred in historical periods, determine all characteristics that caused the accident based on the operating data, and then, for each characteristic, acquire several first evaluation results representing the importance of the characteristics in the current period based on expert knowledge, and several second evaluation results representing the importance of the characteristics in historical periods, and evaluate the reliability of each evaluation result according to the source of each evaluation result to obtain the comprehensive importance of the characteristics, so as to generate an accident correlation index representing the comprehensive correlation between each characteristic and the target hydrogenation unit accident; the potential consequence index acquisition module 22 is implemented according to the above... The method described in step S120 is configured to analyze the types and ranges of damage areas corresponding to different types of accidents in hydrogenation units, as well as the impact of emergency measures on the severity of the consequences, in order to generate a potential consequence index representing the severity of the accident consequences of the target hydrogenation unit; the accident situation development index acquisition module 23 is implemented according to the method described in step S130 above, configured to acquire the change characteristics of working data to predict the working data of the target hydrogenation unit, and then combine the fuzzy comprehensive evaluation method to generate an accident situation development index representing the accident development trend of the target hydrogenation unit; the emergency response level determination module 24 is implemented according to the method described in step S140 above, configured to obtain the current emergency response level of the target hydrogenation unit based on each index.

[0110] This invention discloses a method and system for determining the emergency response level of a hydrogen refueling unit. The method first collects data related to accidents occurring in hydrogen refueling units during historical periods (consequences, emergency measures taken, and unit operating data). Next, based on the operating data, all characteristics causing the accident are identified. Then, for each characteristic, evaluation results based on expert knowledge are obtained, representing the importance of the characteristic in the current and historical periods respectively. The reliability of each evaluation result is evaluated based on its source to obtain a comprehensive characteristic importance, generating an accident correlation index. Then, the types and extent of damage corresponding to different types of hydrogen refueling unit accidents are analyzed, as well as the impact of emergency measures on the severity of consequences, to generate a potential consequence index representing the severity of the accident consequences of the target hydrogen refueling unit. Next, the changing characteristics of the operating data are acquired to predict the operating data of the target hydrogen refueling unit. Then, combined with a fuzzy comprehensive evaluation method, an accident situation development index representing the development trend of the accident in the target hydrogen refueling unit is generated. Finally, the emergency response level of the current target hydrogen refueling unit is obtained based on each index. This invention effectively combines the advantages of expert knowledge and deep learning algorithms, comprehensively considering the correlation between features and accidents, potential consequences of accidents, and the development of accident situations. It integrates qualitative and quantitative research methods to accurately obtain the emergency response level. Furthermore, this invention overcomes the uncertainty and diversity of the characteristics causing accidents in hydrogen refueling units, effectively avoiding randomness and ambiguity in the emergency response process. In addition, this invention can be integrated into intelligent digital management platforms and comprehensive emergency response platforms to achieve accident emergency trigger identification functions, facilitating the linkage between these platforms.

[0111] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0112] Of course, the present invention may have other various embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art can make various corresponding changes and modifications according to the present invention, but these corresponding changes and modifications should all fall within the protection scope of the claims of the present invention.

[0113] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device, or fabricating them separately as individual integrated circuit modules, or fabricating multiple modules or steps as a single integrated circuit module. Thus, the present invention is not limited to any particular hardware and software combination.

[0114] While the embodiments disclosed in this invention are as described above, the content is merely for the purpose of facilitating understanding of the invention and is not intended to limit the invention. Any person skilled in the art to which this invention pertains may make any modifications and variations in form and detail of the implementation without departing from the spirit and scope disclosed herein; however, the scope of patent protection for this invention shall still be determined by the scope defined in the appended claims.

Claims

1. A method for determining the emergency response level of a hydrogenation unit, characterized in that, include: Collect the consequences, emergency measures and operational data of hydrogen refueling unit accidents that occurred in historical periods. Based on the operational data, determine all the characteristics that caused the accident. Then, for each characteristic, obtain several first evaluation results based on expert knowledge that represent the importance of the characteristics in the current period, and several second evaluation results that represent the importance of the characteristics in historical periods. Based on the source of each evaluation result, evaluate the reliability of each evaluation result to obtain the comprehensive importance of the characteristics, so as to generate an accident correlation index that represents the comprehensive correlation between each characteristic and the target hydrogen refueling unit accident. The types and extent of damage corresponding to different types of accidents at hydrogenation units are analyzed, as well as the impact of the emergency measures on the severity of the consequences, in order to generate a potential consequence index representing the severity of the accident consequences of the target hydrogenation unit. The changing characteristics of the working data are obtained to predict the working data of the target hydrogenation unit. Then, combined with the fuzzy comprehensive evaluation method, an accident situation development index representing the accident development trend of the target hydrogenation unit is generated. The emergency response level of the current target hydrogen refueling unit is determined based on each index.

2. The emergency response level determination method according to claim 1, characterized in that, The first evaluation result and the second evaluation result are considered important, unimportant, or their importance cannot be determined. The step of evaluating the reliability of each evaluation result based on its source to obtain the overall importance of the feature includes: Assign a score to each evaluation result according to its importance; For each feature, the score value of each first evaluation result is recorded as the first score value; For each feature, the second evaluation results made by experts with experience in evaluating accident influencing factors are extracted from the plurality of second evaluation results. The score value of each extracted second evaluation result is recorded as the second score value. Then, based on the professional field of the expert who made each extracted second evaluation result, the reliability of each extracted second evaluation result is determined, and a weight is assigned to the corresponding second evaluation result according to the reliability value. Based on this, the first score value and the second score value are combined to calculate the total score value to characterize the comprehensive importance of each feature to the accident of the target hydrogenation unit.

3. The emergency response level determination method according to claim 2, characterized in that, The step of generating an accident correlation index that represents the overall relevance of each feature to an accident at the target hydrogenation unit includes: Using the working data, the correlation between each feature and accidents in the hydrogenation unit is analyzed. Furthermore, the ratio between the number of related accidents and the number of unrelated accidents is calculated for each feature, and this ratio is used as an accident correlation correction factor to correct the accident correlation index. Based on this, combined with the total score, the accident correlation index is obtained.

4. The emergency response level determination method according to claim 3, characterized in that, The total score is calculated using the following expression: Among them, ZS i Let f represent the total score, N represent the total number of first evaluation results, x represent the sequence number of the first evaluation result, and f represent the total score. d b1 and b2 represent coefficients, m represents the total number of first and second evaluation results, n represents the number of evaluation results that are the same as the actual evaluation result, and f represents the first rating value. j represents the second score, w represents the weight, and i represents the feature number.

5. The emergency response level determination method according to claim 4, characterized in that, The accident correlation index is obtained using the following expression: Where S1 represents the accident correlation index, MS i Indicating the degree of relevance, C i This represents the accident-related correction factor, where M represents the total number of all features that cause the accident.

6. The method for determining the emergency response level according to any one of claims 1 to 5, characterized in that, The different accident types of hydrogenation equipment include, but are not limited to: fire, explosion, and poisoning; the types of injury areas include, but are not limited to: death area, serious injury area, and minor injury area; wherein, the step of analyzing the injury area type and injury range corresponding to the different accident types of hydrogenation equipment includes: For each type of hydrogen refueling unit accident, the environmental information of the location of the hydrogen refueling unit when the accident occurred in historical periods is used to obtain the damage range corresponding to different damage area types.

7. The emergency response level determination method according to claim 6, characterized in that, The step of analyzing the impact of the emergency measures on the severity of the consequences includes: Based on the AHP algorithm, a hierarchical model is constructed. By analyzing the first relative importance of each first evaluation indicator relative to the severity of the consequences, and the second relative importance of each second evaluation indicator relative to each of the first evaluation indicators, the impact of each emergency measure on the severity of the consequences is evaluated. This yields a first weight for evaluating the first relative importance of each first evaluation indicator and a second weight for evaluating the second relative importance of each second evaluation indicator. This results in a weight vector for evaluating the relative importance of each second evaluation indicator relative to the severity of the consequences, thus obtaining weight coefficients representing the impact of each evaluation indicator on the severity of the accident consequences of the target hydrogenation unit. Use the severity of consequences as the target layer element; The process control, material isolation, emergency fire protection, and regional planning and fire and explosion prevention in the emergency measures are used as the criteria layer elements and are denoted as the first evaluation index. The sub-emergency measures belonging to the process control, material isolation, emergency fire protection, and regional planning and fire and explosion prevention are respectively used as measure layer elements and are denoted as the second evaluation index.

8. The method for determining the emergency response level according to claim 7, characterized in that, The step of generating a potential consequences index representing the severity of an accident at a target hydrogenation unit includes: The effectiveness of the corresponding emergency measures is determined based on the consequences described, and effective emergency measures and ineffective emergency measures are assigned values ​​of 1 and 0, respectively. For each second evaluation indicator, calculate the product of the emergency measure assignment and the weight coefficient of the second evaluation indicator. For each first evaluation indicator, calculate the sum of the product results of all second evaluation indicators belonging to the first evaluation indicator, and record it as the first result. For each first evaluation index, the product of the sum of the first results and the weight coefficient of the first evaluation index is calculated and recorded as the second result. The sum of all second results is then obtained. Based on this, a corresponding damage degree score is configured for each damage area type according to the degree of damage, and the potential consequence index is obtained by further combining the corresponding damage range.

9. The method for determining the emergency response level according to any one of claims 1 to 8, characterized in that, The steps in generating an accident situation development index that represents the accident development trend of a target hydrogenation unit include: For different possible degrees of accident occurrence at the target hydrogenation unit, corresponding working data ranges are configured. Using the fuzzy comprehensive evaluation method, the correlation between the predicted working data and the predicted membership state representing the probability of an accident at the target hydrogenation unit is established, and the membership function used to predict the probability of an accident at the target hydrogenation unit is obtained. Based on this, the numerical values ​​of the predicted membership state are obtained using the numerical values ​​of the predicted working data, and then the accident situation development index is obtained.

10. The emergency response level determination method according to claim 9, characterized in that, The membership function is represented by the following expression: μ(x) = 0, x ∈ [l1, μ1] Where μ(x) represents the general state membership function, x represents the prediction working data, and μ 1l μ represents the membership function of the lower limit state of level I. 1h μ represents the membership function of the Level I high-limit state. 2l μ represents the membership function of the lower limit state of level II. 2h This represents the membership function for the Level II high-limit state.

11. The method for determining the emergency response level according to any one of claims 1 to 10, characterized in that, The step of acquiring the variation characteristics of the operating data to predict the operating data of the target hydrogenation unit includes: Based on the work data, the work data of the previous period is used as the input information of the preset model, and the work data of the next period is used as the output information of the preset model, thereby constructing a work data prediction model by training the preset model. The current operating data of the target hydrogenation unit is used as the input information of the operating data prediction model to predict the operating data of the target hydrogenation unit, thereby obtaining the corresponding predicted operating data.

12. The method for determining the emergency response level according to any one of claims 2 to 5, characterized in that, The process of determining the reliability of each extracted second evaluation result based on the professional field of the expert who made each extracted second evaluation result includes: If the field of expertise is equipment, then the reliability of the extracted second evaluation result is determined to be the highest; If the field of expertise is emergency response, the reliability of the extracted second evaluation result is determined to be moderate. If the field of expertise is neither equipment nor emergency response, then the reliability of the extracted second evaluation result is determined to be the lowest.

13. The method for determining the emergency response level according to claim 12, characterized in that, The process of assigning weights to the corresponding second evaluation results includes: The second evaluation result, which has the highest reliability, is assigned a weight of 2. The second evaluation result with moderate reliability is assigned a weight of 1.5; The second evaluation result with the lowest reliability is assigned a weight of 1.

14. The emergency response level determination method according to claim 13, characterized in that, The process of assigning score values ​​to each evaluation result according to its importance includes: For evaluation results that are important, unimportant, and uncertain in terms of importance, a score of 1, 0.5, and 0 points will be assigned respectively.

15. An emergency response level determination system for a hydrogenation unit, characterized in that, The emergency response level determination system includes the following modules: The accident correlation index acquisition module is used to collect the consequences, emergency measures and operating data of hydrogen refueling unit accidents in historical periods. Based on the operating data, it determines all the characteristics that caused the accident. Then, for each characteristic, it acquires several first evaluation results based on expert knowledge that indicate the importance of the characteristics in the current period, and several second evaluation results that indicate the importance of the characteristics in historical periods. Based on the source of each evaluation result, it evaluates the reliability of each evaluation result to obtain the comprehensive importance of the characteristics, so as to generate an accident correlation index that indicates the comprehensive correlation between each characteristic and the target hydrogen refueling unit accident. The potential consequences index acquisition module is used to analyze the damage area type and damage range corresponding to different hydrogenation unit accident types, as well as the impact of the emergency measures on the severity of the consequences, in order to generate a potential consequences index representing the severity of the accident consequences of the target hydrogenation unit. The accident situation development index acquisition module is used to acquire the change characteristics of the working data in order to predict the working data of the target hydrogenation unit, and then combine the fuzzy comprehensive evaluation method to generate an accident situation development index that represents the accident development trend of the target hydrogenation unit. The emergency response level determination module is used to determine the emergency response level of the current target hydrogenation unit based on various indices.

Images