Method, system, device and storage medium for analyzing risk of dangerous chemical storage accident

Abstract

This invention relates to the field of hazardous chemical safety management technology, specifically providing a method for risk analysis of hazardous chemical storage accidents. The method includes the following steps: collecting and preprocessing time-series and spatial topology data of the storage area; inputting the time-series data into a time-dynamic risk prediction model, extracting time-sensitive features to output a time-series curve of risk level labeled with risky storage units; labeling high-risk time periods in the curves; constructing a spatial diffusion risk assessment model based on spatial topology data, extracting risk diffusion characteristics of target storage units and associating them with adjacent units, and outputting dynamic risk spatial distribution results; performing spatiotemporal coupling analysis on high-risk time periods and spatial distribution, correcting the results, and outputting early warning information; and finally, calling a management rule base to generate risk management decision instructions. This method covers both spatiotemporal dimensions, achieving dynamic risk assessment, early warning, and management, effectively improving the comprehensiveness and accuracy of hazardous chemical storage risk analysis and management, and providing scientific support for the safety management of storage areas.

Description

Technical Field

[0001] This invention relates to the field of hazardous chemical safety management technology, specifically to a method and system for analyzing the risks of hazardous chemical storage accidents. Background Technology

[0002] Hazardous chemicals (such as gasoline, diesel, and corrosive chemicals) are flammable, explosive, toxic, and harmful. Their storage is a core scenario for safety management. The operating status, environmental changes, and spatial layout of storage units (tanks, warehouses) can all lead to major accidents such as leaks, explosions, and poisoning. It is necessary to identify risks in advance and formulate control measures through scientific risk analysis methods to avoid accidents or reduce consequences.

[0003] Currently, the mainstream analysis methods in the industry are mainly divided into two categories: one is the traditional static assessment method represented by Hazard and Operability Analysis (HAZOP) and static risk matrix. By pre-setting risk factors (such as the toxicity of the storage medium and the pressure of the storage tank) and fixed weights, the risk level of the storage conditions at a specific time is determined, and a static result of "high / medium / low" is output. The other is a single-dimensional dynamic method, such as predicting the time-series changes of risk of a single storage tank through conventional LSTM or simulating the diffusion range of a single leakage space through CFD. This type of method only focuses on a single dimension of "time" or "space" and does not form a linkage.

[0004] While existing methods have some applications, they are limited by a "static and isolated" design logic, making them unsuitable for the "dynamic evolution and spatiotemporal linkage" characteristics of risks in real-world storage scenarios. In terms of time, outputting results based solely on operational data at a single moment ignores the risk accumulation patterns caused by differences in day and night inspections and seasonal humidity changes, failing to capture the "time-sensitive effect" and easily leading to misjudgments. In terms of space, for storage areas with multi-unit cluster layouts, assessing the risk of individual units in isolation ignores the chain reaction of hazardous substance diffusion, resulting in "fragmented" results. At the spatiotemporal coordination level, separating time and space analysis fails to output "when + where" coupled risk information, leading to delayed early warnings, insufficient targeted control, and difficulty in effectively preventing accidents.

[0005] In summary, the field of hazardous chemical storage accident risk analysis urgently needs technical solutions that break through the limitations of static assessment and isolated dimensions to address the aforementioned issues. Summary of the Invention

[0006] To overcome the shortcomings of existing technologies, this invention provides a method and system for risk analysis of hazardous chemical storage accidents, in order to solve the problems in existing technologies.

[0007] One embodiment of the present invention provides a method for risk analysis of hazardous chemical storage accidents, comprising the following steps:

[0008] Collect and preprocess multi-source data on hazardous chemical storage in the storage area, including time-series data and spatial topology data.

[0009] The preprocessed time series data is input into a pre-built time dynamic risk prediction model. The time-sensitive features in the time series data are extracted through the time dynamic risk prediction model. Based on the time-sensitive features, a risk level time series curve for a future preset duration is output, and storage units in the risk level time series curve that are within the preset risk threshold range are associated and labeled.

[0010] The high-risk time period window in the risk level time series curve is marked to indicate that the risk level exceeds the preset safety threshold.

[0011] Based on the preprocessed spatial topology data, a spatial diffusion risk assessment model is constructed. When the risk level time series curve is associated with a storage unit within the preset risk threshold range, the storage unit is taken as the target storage unit. The risk diffusion characteristics of the target storage unit are extracted through the spatial diffusion risk assessment model and associated with adjacent storage units, and the dynamic risk spatial distribution results of the storage area are output.

[0012] The high-risk time window is coupled with the dynamic risk spatial distribution results for spatiotemporal analysis. The analysis results are corrected based on the preset risk association rules, and spatiotemporal coupled risk warning information is output.

[0013] Based on the spatiotemporal coupled risk warning information, a preset risk control rule base is invoked to generate risk control decision instructions for hazardous chemical storage areas.

[0014] By adopting the above scheme, the time-series curves of risk levels with risk storage unit labels and high-risk periods are output by the time dynamic risk prediction model. This can capture the time evolution pattern of risk and solve the problems of static assessment being difficult to reflect dynamic accumulation and prone to misjudgment. Then, the spatial diffusion risk assessment model is used to analyze the diffusion trend of risk storage units and associate it with adjacent storage units to cover the overall risk of the storage area and solve the problems of isolated assessment ignoring chain effects and fragmented results. Finally, the spatiotemporal coupling analysis outputs early warning and generates control instructions to achieve spatiotemporal risk coordination. This solves the problems of spatiotemporal fragmentation and delayed early warning in existing methods, and improves the dynamism, comprehensiveness and control accuracy of hazardous chemical storage risk analysis, providing scientific support for safety management.

[0015] In one embodiment, the construction of the time-dynamic risk prediction model specifically includes the following steps:

[0016] Historical time-series sample data, including storage unit operating condition data, environmental dynamic data, and human intervention data, are acquired from the storage area. The historical time-series sample data is preprocessed to obtain a standardized time-series feature set. Based on historical accident data of the same type of storage unit within a preset time period, risk level labels are marked for the standardized time-series feature set. Data before the accident is marked as high risk, and normal operation data is marked as low risk.

[0017] Construct a time-series prediction model that includes a feature extraction layer for extracting time-sensitive features and a prediction output layer for outputting time-series curves of risk levels;

[0018] A standardized time-series feature set labeled with risk level is input into the time-series prediction model architecture, and the model parameters are adjusted using a preset training strategy to obtain a trained time-dynamic risk prediction model.

[0019] The preprocessed time series data is input into the trained time dynamic risk prediction model. The time-sensitive features in the time series data are extracted through the feature extraction layer to form a sensitive feature vector set that includes storage unit operating condition features, environmental dynamic features, and human intervention features. The sensitive feature vector set is then input into the prediction output layer, which outputs the risk level time series curve for a future preset duration.

[0020] By adopting the above approach, the training foundation is built using three types of historical time-series sample data and risk level labels, which can solve the prediction bias problems caused by one-sided training data and ambiguous labels. Then, a long short-term memory network architecture with an attention mechanism is used to extract key time-sensitive features and control the training accuracy, solving the problems of missing key features and overfitting failure in conventional time-series models. Finally, by outputting a structured sensitive feature vector set, the problems of black box feature extraction and spatiotemporal coupling input bias are solved, thereby improving the accuracy, reliability and scenario adaptability of time-dynamic risk prediction, and providing more refined time dimension support for hazardous chemical storage risk analysis.

[0021] In one embodiment, the step of inputting the sensitive feature vector set into the prediction output layer and outputting a risk level time series curve for a preset future duration through the prediction output layer specifically includes the following steps:

[0022] The sensitive feature vector set is associated with the hazardous chemical type attribute of the corresponding storage unit to form an associated dataset; based on the associated dataset, the sensitive feature vector set is grouped according to the hazardous chemical type, and the sensitive feature vectors in each group are sorted by time series to form a time series feature matrix hierarchically according to the risk characteristics of hazardous chemicals.

[0023] The dynamic correlation between adjacent vectors in each hierarchical time-series feature matrix is ​​analyzed by a preset time-series correlation algorithm. Based on the attention mechanism, a preset high-impact weight value is assigned to the accident-related feature vectors that match the historical accident features, and a preset basic weight value is assigned to other non-accident-related auxiliary feature vectors. The historical accident features are extracted based on historical accident data of the same type of storage unit within a preset time period.

[0024] Based on the preset risk level mapping rules for hazardous chemical types, combined with the risk level trend correction factor, and integrating the high impact weight value of the accident-related feature vector with the basic weight value of the auxiliary feature vector, the risk level value for each period within the preset time period is calculated.

[0025] The risk level values ​​for each time period are arranged continuously along the time axis to generate a risk level time series curve. The risk level time series curve includes storage units marked with spatial association unit labels that are within the preset risk threshold range.

[0026] By adopting the above scheme, a hierarchical matrix is ​​formed by associating sensitive feature vector sets with hazardous chemical type attributes, grouping and sorting them by time series. This solves the problems of mixed analysis of different hazardous chemical features and masking of key risk characteristics in the existing prediction output layer. Furthermore, a time series correlation algorithm is used to analyze dynamic correlations, and an attention mechanism is used to differentiate the weighting of accident-related feature vectors and auxiliary feature vectors. Combined with hazardous chemical type-specific mapping rules and time series trend correction factors, risk levels are calculated, addressing the issues of subjective weight settings, undifferentiated risk calculation standards, and missed detection of gradual risks in the conventional output layer. Finally, by generating risk level time series curves with labels for high-risk storage units and spatially related units, the problem of insufficient targeted prevention and control due to prediction output only reflecting time-dimensional risk and ignoring spatial chain effects is solved. This comprehensively improves the accuracy, information completeness, and prevention and control guidance of the risk level time series curve output, providing more targeted technical support for the prediction output stage of hazardous chemical storage risk analysis.

[0027] In one embodiment, the method for determining the preset risk threshold includes the following steps:

[0028] Based on the hazardous chemical type attribute of the storage unit, the basic safety threshold of the corresponding type of hazardous chemical is retrieved as the initial risk threshold;

[0029] The risk level threshold of the corresponding type of hazardous chemicals at the time of the historical accident is extracted from the characteristics of the historical accident. The initial risk threshold is corrected based on the risk level threshold to obtain the type-specific benchmark threshold.

[0030] Real-time dynamic impact factors are extracted from the set of sensitive feature vectors. Based on the extracted real-time dynamic impact factors, the type-specific benchmark threshold is dynamically adjusted to determine the preset risk threshold of the corresponding storage unit.

[0031] By adopting this scheme, using the basic safety threshold corresponding to the type and attribute of hazardous chemicals as the initial standard, the problems of misjudgment caused by the disconnect between existing preset risk thresholds and the characteristics of hazardous chemicals, and their excessive generality can be solved. Furthermore, by combining the historical accident risk threshold correction benchmark of the same type of hazardous chemicals, the problem of threshold setting lacking actual accident data support can be solved. Finally, the dynamic influencing factors in the sensitive feature vector set are introduced for real-time adjustment, solving the problem that static thresholds cannot adapt to real-time operating condition changes. Overall, the pertinence, scientificity, and dynamic adaptability of preset risk thresholds are improved, providing an accurate basis for judging high-risk storage units for the risk level time series curve.

[0032] In one embodiment, the step of constructing a spatial diffusion risk assessment model based on preprocessed spatial topology data, where a storage unit within a preset risk threshold range is associated with a risk level time series curve, and the storage unit is taken as the target storage unit, the risk diffusion characteristics of the target storage unit are extracted through the spatial diffusion risk assessment model and associated with adjacent storage units, and the dynamic risk spatial distribution result of the storage area is output, specifically includes the following steps:

[0033] The preprocessed spatial topology data is analyzed hierarchically to extract spatial correlation parameters and physical barrier levels between the target storage unit and adjacent storage units. Based on the diffusion characteristic parameters of hazardous chemicals within the target storage unit, a spatial diffusion risk assessment model is constructed.

[0034] When the risk level time series curve is associated with a storage unit within the preset risk threshold range, the real-time risk level value of the target storage unit is obtained and input into the spatial diffusion risk assessment model. The risk diffusion characteristics of the target storage unit are extracted through the spatial diffusion risk assessment model. The risk diffusion characteristics include the risk diffusion rate, diffusion radius and hazardous substance concentration decay coefficient of the target storage unit.

[0035] Based on the extracted risk diffusion characteristics and the spatial correlation parameters between the target storage unit and adjacent storage units, the risk diffusion characteristics of the target storage unit are associated with the corresponding adjacent storage units according to the preset correlation rules. The risk level of each adjacent storage unit affected by diffusion is calculated, and a correlation table containing the target storage unit, adjacent storage units, diffusion path and affected risk level is generated.

[0036] Based on the risk change trend of the risk level time series curve, the correlation table is overlaid and matched with the spatial map of the storage area to generate and output the dynamic risk spatial distribution result of the storage area that is dynamically updated over time. The dynamic risk spatial distribution result includes the actual risk level and diffusion impact range of each storage unit at different time periods.

[0037] By adopting the above scheme, a spatial diffusion risk assessment model coupling spatial correlation and diffusion characteristics is constructed using hierarchical analysis of spatial topology data. Risk diffusion rate, radius, and concentration decay coefficient are extracted as quantitative features. This addresses the problems of existing spatial diffusion models being too general (relying only on unit distance, ignoring connectivity and physical barriers) and having abstract diffusion characteristics (lacking clear quantitative parameters), leading to a weak assessment foundation. Furthermore, adjacent units are correlated according to a diffusion range coverage priority + connectivity influence weighting rule, and the affected risk level is calculated using the concentration decay coefficient. This solves the problems of inaccurate correlation and level deviation caused by the coarse correlation of adjacent units (selected only by distance) and the lack of quantitative basis for risk level calculation. Finally, the spatial distribution results are dynamically updated using risk level time-series curves, addressing the problem that existing static spatial assessments cannot adapt to real-time changes in operating conditions (such as risk escalation / decay), resulting in lagging dynamic control. Overall, this improves the accuracy, dynamic adaptability, and information completeness of the dynamic risk spatial distribution of storage areas, providing precise spatial dimension support for high-risk area positioning, emergency resource allocation, and real-time prevention and control strategy formulation.

[0038] In one embodiment, the step of extracting the risk diffusion characteristics of the target storage unit through the spatial diffusion risk assessment model further includes the following step:

[0039] Obtain the physical properties of the target storage unit and the hazardous chemical type properties inside it, and determine whether the target storage unit belongs to a special scenario based on a preset judgment threshold;

[0040] When the target storage unit belongs to a special scenario, a preset special scenario correction coefficient table is called to match the correction coefficient corresponding to the target storage unit. The special scenario correction coefficient table contains correction coefficients corresponding to several special scenarios.

[0041] The extracted risk diffusion characteristics are dynamically corrected based on the correction coefficient corresponding to the target storage unit, resulting in the corrected risk diffusion characteristics.

[0042] By adopting the above scheme, based on the physical properties of the target storage unit and the type of hazardous chemicals, and combined with preset judgment thresholds to identify special scenarios, the problem of insufficient scenario adaptability caused by the lack of clear definition of special scenarios (such as low temperature / high pressure units, flammable / highly corrosive hazardous chemicals) and the assessment based solely on general rules in existing spatial diffusion assessments can be solved. Furthermore, by calling the preset special scenario correction coefficient table to match the corresponding correction coefficients, the problem of lack of quantitative basis for risk diffusion characteristic correction in special scenarios and the subjectivity of correction parameters can be solved. Finally, the risk diffusion characteristics are dynamically corrected according to the matched correction coefficients, solving the problem of assessment distortion caused by using the same diffusion characteristic parameters for special scenarios and conventional scenarios (such as overestimation of the diffusion range of low temperature units and misjudgment of the concentration decay of flammable hazardous chemicals) can be solved. Overall, the accuracy of risk diffusion characteristic extraction and scenario adaptability in special scenarios are improved, providing more practical technical support for the risk assessment of hazardous chemical storage space diffusion, and avoiding risk assessment deviations and subsequent management errors caused by misjudgments in special scenarios.

[0043] In one embodiment, the step of performing spatiotemporal coupling analysis on the high-risk time window and the dynamic risk spatial distribution results, correcting the analysis results based on preset risk association rules, and outputting spatiotemporally coupled risk early warning information specifically includes the following steps:

[0044] Extract the time boundary parameters of the high-risk period window, including the high-risk start time, duration, and peak risk level time;

[0045] The time boundary parameters of the high-risk period window are mapped to the dynamic risk spatial distribution results in a spatiotemporal dimension to match the risk spatial distribution status of the storage area corresponding to each time node, forming an initial spatiotemporal coupling matrix. The initial spatiotemporal coupling matrix contains the associated data of time node, storage unit, risk level, and diffusion range.

[0046] The initial spatiotemporal coupling matrix is ​​modified based on preset risk association rules, which include time priority weighting rules, spatial correlation rules, and risk superposition rules.

[0047] Based on the corrected spatiotemporal coupling matrix, spatiotemporal coupling risk warning information is generated and output.

[0048] By adopting the above scheme, using the time boundary parameters (start time, duration, and peak time) of high-risk period windows as time anchors, the problem of insufficient spatiotemporal correlation accuracy caused by the ambiguous description of existing high-risk periods (only defining the time range without key nodes) can be solved. Furthermore, by constructing an initial coupling matrix containing time-space-risk-diffusion data through spatiotemporal dimension mapping, the problem of weak coupling analysis foundation caused by the fragmentation of time and space risk data (each presented independently without a clear correspondence) can be solved. Finally, the matrix is ​​corrected based on time priority weights, spatial correlation, and risk superposition rules, and early warning information is output. This solves the problem of poor early warning targeting and insufficient control guidance caused by conventional spatiotemporal coupling simply splicing data (without distinguishing risk priority from actual hazard differences). Overall, the accuracy, priority differentiation capability, and early warning practicality of spatiotemporal coupled risk analysis are improved, providing decision support with both time sensitivity and spatial targeting for dynamic risk collaborative management of hazardous chemical storage areas.

[0049] This application also relates to a hazardous chemical storage accident risk analysis system, including:

[0050] The data acquisition module is used to collect and preprocess multi-source data on hazardous chemical storage in the storage area. The multi-source data on hazardous chemical storage includes time-series data and spatial topology data.

[0051] The risk prediction module is used to input pre-processed time series data into a pre-built time dynamic risk prediction model, extract time-sensitive features from the time series data through the time dynamic risk prediction model, output a risk level time series curve for a future preset duration based on the time-sensitive features, and associate and label the storage units in the risk level time series curve that are within the preset risk threshold range.

[0052] The annotation module is used to annotate high-risk time period windows in the risk level time series curve where the risk level exceeds a preset safety threshold;

[0053] The risk assessment module is used to construct a spatial diffusion risk assessment model based on preprocessed spatial topology data. When the risk level time series curve is associated with a storage unit within the preset risk threshold range, the storage unit is taken as the target storage unit. The risk diffusion characteristics of the target storage unit are extracted through the spatial diffusion risk assessment model and associated with adjacent storage units, and the dynamic risk spatial distribution results of the storage area are output.

[0054] The coupling analysis module is used to perform spatiotemporal coupling analysis on high-risk time windows and dynamic risk spatial distribution results, correct the analysis results based on preset risk association rules, and output spatiotemporal coupling risk warning information.

[0055] The decision generation module is used to generate risk control decision instructions for hazardous chemical storage areas by calling a preset risk control rule base based on the spatiotemporal coupled risk warning information.

[0056] This application also relates to a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method for risk analysis of hazardous chemical storage accidents.

[0057] This application also relates to a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described method for risk analysis of hazardous chemical storage accidents.

[0058] The method and system for risk analysis of hazardous chemical storage accidents provided in the above embodiments have the following beneficial effects:

[0059] By using a time-dynamic risk prediction model to output risk level time-series curves with risk storage unit labels and high-risk periods, the evolution of risk over time can be captured, solving the problems of static assessments failing to reflect dynamic accumulation and being prone to misjudgment. Then, a spatial diffusion risk assessment model is used to analyze the diffusion trend of risk storage units and correlate it with adjacent storage units, thus covering the overall risk of the storage area and solving the problems of isolated assessments ignoring chain effects and fragmented results. Finally, spatiotemporal coupling analysis outputs early warnings and generates control instructions, achieving spatiotemporal risk synergy. This solves the problems of spatiotemporal fragmentation and delayed early warning in existing methods, comprehensively improving the dynamism, comprehensiveness, and control accuracy of hazardous chemical storage risk analysis, providing scientific support for safety management. Attached Figure Description

[0060] Figure 1 A flowchart of a method for risk analysis of hazardous chemical storage accidents provided in an embodiment of the present invention;

[0061] Figure 2 This is a schematic block diagram of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0062] The technical solutions in the embodiments of the present invention will now be clearly and completely described in conjunction with the accompanying drawings.

[0063] Reference Figure 1 One embodiment of the present invention provides a method for risk analysis of hazardous chemical storage accidents, comprising the following steps:

[0064] S10. Collect and preprocess multi-source data of hazardous chemical storage in the storage area. The multi-source data of hazardous chemical storage includes time series data and spatial topology data.

[0065] In this embodiment, the core objective is to provide complete and standardized foundational data for subsequent time-based dynamic risk prediction and spatial diffusion risk assessment, avoiding analytical biases caused by missing data or inconsistent formats. For data acquisition, considering the explosion-proof and corrosion-resistant requirements of hazardous chemical storage areas, explosion-proof IoT sensor arrays (such as flameproof temperature sensors, intrinsically safe pressure transmitters, and electrochemical gas concentration sensors) are used to collect time-series data in real time. The core data includes storage unit operating parameters (such as temperature, pressure, and liquid level), environmental parameters (such as temperature and humidity, and toxic / flammable gas concentrations), and personnel inspection records. Spatial topology data is obtained through CAD drawing analysis or on-site distance measurement, and the core data includes storage unit locations, unit spacing, connectivity and separation relationships such as pipelines / fire dikes, and the distribution of surrounding sensitive targets (such as water sources and residential areas). For data preprocessing, the time-series data undergoes routine noise reduction, missing value completion, and format standardization (which can be achieved using conventional methods by those skilled in the art); duplicate records are removed from the spatial topology data, and the data is standardized into a format suitable for spatial analysis. The preprocessed time series data is used for subsequent time dynamic risk prediction, and the preprocessed spatial topology data is used for subsequent spatial diffusion risk assessment.

[0066] S20. Input the preprocessed time series data into the pre-constructed time dynamic risk prediction model, extract the time-sensitive features in the time series data through the time dynamic risk prediction model, output the risk level time series curve for a future preset duration based on the time-sensitive features, and associate and label the storage units in the risk level time series curve that are within the preset risk threshold range.

[0067] In this embodiment, the core objective is to achieve dynamic prediction of future risks in the storage area based on standardized time-series data. This provides a time-dimensional risk benchmark for subsequent high-risk period identification and spatial diffusion analysis, avoiding the risk prediction lag caused by relying solely on historical static data. For input data, the preprocessed time-series data (including storage unit operating parameters, environmental parameters, and personnel inspection records, etc.) from S10 is used. Data format verification ensures that it matches the input interface of the time-dynamic risk prediction model, allowing for direct use without secondary conversion.

[0068] When using this model to automatically analyze time-series data, the focus is on extracting time-sensitive features strongly correlated with risk evolution (such as hourly upward trends in temperature parameters, abnormal fluctuation frequencies of gas concentrations, and durations of pressure exceedances). Based on these extracted features, the model outputs a risk level time-series curve for a preset duration (the preset duration can be flexibly set according to actual management needs, such as 24 hours for daily monitoring and 72 hours for major holidays). This curve contains two key pieces of information: first, the overall risk level of the storage area at each time point (reflecting the overall safety status of the area); and second, the individual risk level of each storage unit at each time point (locating specific high-risk units). For example, if the preset duration is 24 hours and the time points are divided into 1-hour intervals, the curve will show that "at 10:00, the overall risk level of the area is Level II (medium risk), with Unit A (storing ethanol) at Level III (higher risk) and Unit B (storing sodium hydroxide) at Level I (low risk); at 15:00, the overall risk level of the area rises to Level III, Unit A reaches Level IV (high risk), and Unit B remains at Level I."

[0069] Meanwhile, based on preset risk thresholds (such as setting "risk level ≥ III" as the high-risk judgment standard), storage units with excessive risk levels are marked in the risk level time series curve with special markers (such as red dots and bold lines), clearly showing the distribution characteristics of high-risk units at different time points.

[0070] It should be noted that the rules for determining risk levels and the preset thresholds (such as the quantitative indicator boundaries of levels I to IV, the high-risk determination standard "≥ level III", etc.) can be flexibly set by those skilled in the art according to the actual operation of the storage area (such as the type of hazardous chemicals, the distribution of surrounding sensitive targets, and the enterprise's safety management requirements). Such adaptive adjustments are conventional technical means and are not rigidly limited here.

[0071] The output risk level time series curves and high-risk unit annotation information will be directly used in the subsequent S30 (high-risk period window annotation) to "lock the time interval of high risk concentration" and in S40 (spatial diffusion risk assessment) to "determine the target storage unit that needs to be analyzed", providing a unified time dimension benchmark for cross-stage risk analysis.

[0072] S30. Mark the high-risk period window in the risk level time series curve where the risk level exceeds the preset safety threshold.

[0073] In this embodiment, the core objective is to accurately pinpoint the time intervals of risk exceeding the standard from the risk level time series curve, providing clear high-risk time boundaries for subsequent spatiotemporal coupling analysis and avoiding insufficient targeting of risk control due to ambiguity in the time dimension. For input data, the risk level time series curve (including overall and individual risk levels) output by S20 is used, combined with a preset safety threshold as the judgment benchmark.

[0074] In practice, the risk level change trend of the time series curve is used to identify time intervals that continuously or intermittently exceed the preset safety threshold. If the risk level of a single time node exceeds the standard, and adjacent nodes (such as within one hour before or after) continue to exceed the standard, they are merged into a continuous high-risk time period window. If there are multiple independent exceeding intervals (such as 9-11 am and 3-5 pm), they are marked as independent windows respectively. For example, based on the preset threshold of "risk level ≥ III", two high-risk time period windows are identified from the 24-hour time series curve output by S20: "8:00-10:30 (region as a whole and unit A continuously at level III)" and "14:00-16:00 (region as a whole at level II but unit A reaches level IV)". Each window is marked with the start time, end time, and the core exceeding unit within the corresponding time period. The output high-risk time period window information will directly serve as the key input for the time dimension mapping in S50 (spatiotemporal coupling analysis), used to match the dynamic risk spatial distribution state of the corresponding time period, providing time coordinate support for accurately locating high-risk spatiotemporal overlapping areas.

[0075] S40. Based on the preprocessed spatial topology data, construct a spatial diffusion risk assessment model. When the risk level time series curve is associated with a storage unit within the preset risk threshold range, take the storage unit as the target storage unit, extract the risk diffusion characteristics of the target storage unit through the spatial diffusion risk assessment model and associate it with adjacent storage units, and output the dynamic risk spatial distribution results of the storage area.

[0076] In this embodiment, the core objective is to transform high-risk units identified in the time dimension into spatial diffusion risk analysis, constructing a "unit risk-spatial diffusion" correlation logic. This provides a spatial risk distribution benchmark for subsequent spatiotemporal coupling analysis, avoiding omissions in the prevention and control scope caused by focusing only on the risk of a single unit and ignoring the surrounding chain effects. For the input data, on the one hand, the preprocessed spatial topology data in S10 (including storage unit locations, unit spacing, connectivity and separation relationships such as pipes / fire dikes, and distribution of surrounding sensitive targets) is used as the spatial basis for model construction. On the other hand, combined with the risk level time series curve output in S20, storage units marked within the preset risk threshold range (such as units A and C with risk levels ≥ III) are used as target storage units to clarify the analysis object.

[0077] The specific operation consists of two steps: First, a spatial diffusion risk assessment model is constructed based on spatial topology data. The model integrates unit physical attributes (such as the type of protective facilities), hazardous chemical diffusion characteristics (such as evaporation rate), and spatial correlations (such as unit spacing and barrier conditions) to provide a computational framework for diffusion feature extraction. Second, the model extracts the risk diffusion characteristics of the target storage unit (such as diffusion rate, diffusion radius, and hazardous substance concentration decay coefficient), and associates these characteristics with adjacent storage units according to the rule of "prioritizing diffusion range coverage + weighting the impact of connectivity relationship" (for example, if target unit A is a tank area storing ethanol with a diffusion radius of 50 meters, the model will associate this diffusion characteristic with units B and D within 50 meters that are not blocked by firewalls, and calculate the risk level of units B and D affected based on the concentration decay coefficient).

[0078] The final output, the dynamic spatial distribution of risks in the storage area, uses a spatial map as its carrier. It includes the location of the target storage unit, the risk level of adjacent units affected by diffusion, diffusion path markings (e.g., the diffusion direction of "unit A → unit B"), and the diffusion range boundary. For example, target unit A is marked in red on the map, and units B (risk level II) and D (risk level III) affected by diffusion are marked in yellow, with a 50-meter diffusion range outlined by a dashed line. This result will be directly used for S50 (spatiotemporal coupling analysis) and matched with high-risk time windows (marked in S30) to form a three-dimensional analysis basis of "time-space-risk".

[0079] S50. Perform spatiotemporal coupling analysis on the high-risk time window and the dynamic risk spatial distribution results, correct the analysis results based on the preset risk association rules, and output spatiotemporal coupling risk warning information.

[0080] In this embodiment, the core objective is to break down the disconnect between the time dimension (high-risk periods) and the spatial dimension (diffusion distribution), achieving a three-dimensional collaborative analysis of "when, where, and what kind of risk." This avoids missing key overlapping risks (such as overlapping areas between high-risk periods and high diffusion ranges) due to analyzing the spatiotemporal dimensions in isolation, providing accurate risk location data for subsequent control instructions. For input data, the high-risk period window (including start / end times and core exceeding units) output by S30 and the dynamic risk spatial distribution results (including unit location, diffusion range, and affected risk level) output by S40 are directly invoked, ensuring that the two types of data are accurately correlated through "storage unit ID" and "time node."

[0081] The specific operation is divided into two steps: The first step is spatiotemporal coupling analysis, which maps and matches each high-risk time window with the dynamic risk spatial distribution of the corresponding time node; for example, the 8:00-10:30 (high risk of unit A) time period marked by S30 corresponds to the spatial state of unit A diffusion range of 50 meters, unit B (risk level II affected by diffusion), and unit D (risk level III affected by diffusion) in S40 during the 8:00-10:30 time period, forming an initial coupling matrix of "time-space-risk level", which clearly presents the differences in risk spatial distribution in different time periods. The second step is to correct the analysis results based on preset risk association rules. These preset rules are set around the "risk superposition effect," such as: time priority weighting rules (the spatial risk level is raised by 1 level at the peak of the risk level during a high-risk period), spatial correlation rules (if a unit affected by diffusion is close to sensitive targets such as water sources or residential areas, the risk level is raised even further), and risk superposition rules (if the same unit is continuously in the diffusion influence area during multiple high-risk periods, it is marked as a "key concern unit"). By correcting the rules, "low-priority risks" (such as slight diffusion during non-peak periods) are eliminated, while "high-priority risks" (such as diffusion during peak periods and near sensitive targets) are highlighted.

[0082] The final output of spatiotemporally coupled risk warning information is presented in the form of a visual report or structured data, containing the following core content: First, the high-risk area with spatiotemporal overlap (e.g., "8:30-9:30 (peak of high-risk period) + within 50 meters of Unit A's diffusion range (including Unit D and the water source to the north)"); second, the warning level of the corresponding area (e.g., divided into "Blue Warning - General Attention", "Orange Warning - Key Prevention and Control", and "Red Warning - Emergency Response" according to the degree of risk overlap); and third, risk impact indications (e.g., "The red warning area may affect residential areas within 300 meters, requiring a response within 1 hour"). This warning information will directly serve as the core input for S60 (Risk Control Decision Command Generation), ensuring that subsequent control measures accurately correspond to the "high-risk spatiotemporally overlapped area" and avoiding indiscriminate prevention and control that leads to resource waste.

[0083] S60. Based on the spatiotemporal coupled risk warning information, call the preset risk control rule base to generate risk control decision instructions for the hazardous chemical storage area.

[0084] In this embodiment, the core objective is to transform the "high-risk spatiotemporal overlay information" obtained from spatiotemporal coupling analysis into actionable control measures. This avoids the problem of "warning without action" caused by a disconnect between early warning information and actual control, ensuring that risks can be controlled in a timely and accurate manner. For input data, on the one hand, the spatiotemporal coupling risk early warning information output by the S50 (including spatiotemporally overlaid high-risk areas, warning levels, and risk impact indicators) serves as the core basis for matching control measures. On the other hand, a pre-set risk control rule base is invoked. This rule base is a "warning level - control measure" correspondence system built based on hazardous chemical storage safety regulations and historical accident handling experience. For example, it is categorized by warning level (blue / orange / red) and by risk type (spread risk, unit-specific risk, sensitive target-related risk), ensuring that measures are accurately matched to risk characteristics.

[0085] The specific operation consists of two steps: The first step is rule matching, which calls the rule base based on the key parameters of the spatiotemporally coupled risk warning information. For example, if the warning information is "Red Warning (Unit A, 8:30-9:30, diffusion range includes the water source to the north)," then the control rules corresponding to "Red Warning + Sensitive Target (Water Source) Associated Risk" are matched from the rule base, such as "Immediately activate the emergency ventilation system of Unit A, set up an isolation zone within 50 meters of the diffusion range, and dispatch emergency monitoring personnel to the vicinity of the water source to detect the concentration within 10 minutes." The second step is instruction generation, which transforms the matched rules into structured control decision instructions. The instructions clearly define the control objects (Unit A, the area surrounding the water source), specific measures (ventilation, isolation, personnel dispatch), execution time limit (within 10 minutes, continuing until the warning is lifted), and responsible entities (on-site operation and maintenance team, emergency monitoring team), to avoid execution deviations caused by ambiguous measures.

[0086] The final risk management decision instructions can be synchronized to various implementing departments (such as operations and maintenance and emergency response) through the system platform. These instructions can be in the form of text instructions or visual task lists (marked with a map of the control area and priority of measures). For example, the instruction list might indicate: "Priority 1: The emergency monitoring team must arrive at the water source north of Unit A before 9:00 AM and perform gas concentration detection every 5 minutes; Priority 2: The operations and maintenance team must activate the explosion-proof ventilation equipment in Unit A before 8:40 AM and continue operating it until the warning is lifted." This instruction directly guides on-site safety management operations, achieving closed-loop management of "early warning-response" and minimizing the probability of risk escalation.

[0087] In one embodiment, step S20, the construction of the time-dynamic risk prediction model, specifically includes the following steps:

[0088] S21. Obtain historical time-series sample data in the storage area, including storage unit operating condition data, environmental dynamic data, and human intervention data. Preprocess the historical time-series sample data to obtain a standardized time-series feature set. Based on historical accident data of the same type of storage unit within a preset time period, label the standardized time-series feature set with risk level tags, wherein data before the accident is labeled as high risk and normal operation data is labeled as low risk.

[0089] S22. Construct a time series prediction model that includes a feature extraction layer for extracting time-sensitive features and a prediction output layer for outputting the time series curve of risk level;

[0090] S23. Input the standardized time series feature set labeled with risk level into the time series prediction model architecture, adjust the model parameters using a preset training strategy, and obtain the trained time dynamic risk prediction model.

[0091] S24. Input the preprocessed time series data into the trained time dynamic risk prediction model, extract the time-sensitive features from the time series data through the feature extraction layer, form a sensitive feature vector set containing storage unit operating condition features, environmental dynamic features and human intervention features, and input the sensitive feature vector set into the prediction output layer, and output the risk level time series curve for the future preset duration through the prediction output layer.

[0092] In this embodiment, as described in step S21, the core objective is to provide high-quality historical sample data and labels for training the time-dynamic risk prediction model, ensuring that the model can learn the patterns of risk evolution and avoiding low prediction accuracy due to insufficient training data or ambiguous labels. For input data, three types of historical time-series samples are focused on: storage unit operating condition data (such as equipment parameters like tank pressure, liquid level, and pump operating status), environmental dynamic data (such as real-time temperature, humidity, wind speed, and atmospheric pressure in the storage area), and human intervention data (such as abnormal feedback in inspection records, start-up and shutdown records in operation logs, and maintenance time nodes). The data coverage must include the normal operating cycle and the periods before and after accidents to ensure sample diversity.

[0093] Specifically, the process includes: First, data preprocessing, which involves cleaning the collected historical time-series samples (removing abnormal jumps caused by sensor malfunctions), normalizing (unifying parameters of different magnitudes to the [0,1] range, such as converting temperatures from -20 to 50℃ to standardized values), and time-series alignment (integrating multi-source data by timestamp to form a continuous sequence in minutes), ultimately resulting in a structured standardized time-series feature set. Second, risk level labeling, which assigns labels to the standardized time-series feature set based on historical accident data of similar storage units within a preset time period (such as ethanol tank leaks and sodium hydroxide corrosion accidents). For example, if an ethanol tank leaks on June 10, 2023, its historical data from 00:00 on June 7 to 08:00 on June 10 is labeled as "high risk," and the operational data without abnormalities from May 1 to May 31, 2023 is labeled as "low risk." The labels and feature sets are precisely linked through timestamps. The output standardized time series feature set and corresponding risk level labels will be directly used for training the time dynamic risk prediction model (such as model parameter optimization and feature weight learning), providing corresponding "feature-label" samples for the model to identify risk evolution patterns, and ensuring that the risk level time series curves output by the subsequent model have reliable historical data support.

[0094] As described in step S22, the core objective is to build a model architecture that can accurately capture the temporal evolution of risks, providing a basic framework for time-dynamic risk prediction and avoiding incomplete extraction of time-sensitive features or deviations in prediction output due to unreasonable model structure design. Model construction needs to be combined with the characteristics of the standardized time-series feature set output in S21 (including multi-dimensional time-series data and risk level labels) to ensure that the architecture adapts to the dual needs of "time series analysis + risk level prediction".

[0095] The specific operation focuses on a two-layer core structure design: First, the feature extraction layer, primarily responsible for capturing time-sensitive features related to risk from time-series data. It employs a network structure capable of handling long-term temporal dependencies (such as a Long Short-Term Memory network incorporating attention mechanisms). For example, attention mechanisms assign higher weights to key features like "continuously rising temperature trends" and "abnormal pressure fluctuation frequency," enhancing the ability to extract risk precursor signals. Second, the prediction output layer, based on the key features output by the feature extraction layer, constructs a mapping relationship to generate a risk level time-series curve. A fully connected network combined with temporal smoothing processing (such as sliding window averaging) ensures that the output curve reflects both the dynamic changes in risk over time and clearly distinguishes the individual risk differences between different storage units (e.g., by using multiple output nodes to correspond to the risk levels of each unit). The completed time-series prediction model will serve as the foundational architecture for subsequent model training (e.g., parameter optimization using the standardized time-series feature set and labels from S21), ensuring the model's ability to learn risk evolution patterns from historical data and providing reliable model support for "inputting time-series data and outputting risk level time-series curves" in S20.

[0096] As described in step S23, the core objective is to optimize model parameters using labeled historical sample data, enabling the time-series prediction model to possess accurate risk level prediction capabilities. This avoids prediction biases caused by unreasonable initial values ​​of model parameters (such as missing high-risk periods or mislabeling low-risk units), providing a performance-compliant model tool for the subsequent dynamic risk prediction in S20. For the input data, on the one hand, the "standardized time-series feature set labeled with risk levels" (containing feature data and corresponding high / low risk labels) output from S21 is used; on the other hand, the "time-series prediction model architecture" (containing feature extraction layer and prediction output layer) constructed in S22 is reused to ensure that the data and model input dimensions and formats are completely matched.

[0097] The specific operation consists of two steps: The first step is data partitioning and training initialization. The labeled standardized time-series feature set is divided into a training set (for parameter learning) and a validation set (for performance verification) according to a preset ratio (e.g., 7:3). For example, 70% of the "historical data of ethanol storage tanks (including high-risk labels before accidents and normal low-risk labels)" is used as the training set, and 30% is used as the validation set. At the same time, the model parameters are initialized (e.g., attention weights of the feature extraction layer and connection weights of the prediction output layer). The second step is to adjust the parameters using a preset training strategy. The training strategy is designed around "improving prediction accuracy and avoiding overfitting". For example, the optimizer is Adam (adaptive learning rate optimizer, which can dynamically adjust the learning step size), the loss function is cross-entropy loss (to measure the deviation between the model's predicted label and the true label), and an early stopping strategy is added during training (training is stopped when the prediction error of the validation set does not decrease for 5 consecutive rounds to prevent the model from overfitting the training data). At the same time, the risk level prediction accuracy of the model on the validation set is calculated after each round of training (e.g., accuracy ≥ 90%) until the parameters are adjusted to meet the performance requirements of the validation set.

[0098] The final trained time-dynamic risk prediction model is obtained. This model must meet the core performance indicators (such as the accuracy of risk level prediction on the validation set ≥90% and the false negative rate during high-risk periods ≤5%). It can be directly used in step S20 to input the preprocessed real-time time series data into the model, and it can stably output the risk level time series curve for the future preset duration, providing reliable model support for risk prediction in the time dimension.

[0099] As described in step S24, the core objective is to apply the trained time-dynamic risk prediction model (output of S23) to actual time-series data, achieving a complete implementation from "data input" to "feature extraction" and then to "risk prediction." This avoids a disconnect between the model training and actual data application, ensuring a stable output of time-series curves reflecting the true evolution of risk, and providing a temporal dimension basis for subsequent high-risk identification and spatial analysis. For the input data, on the one hand, the preprocessed real-time time-series data from S10 (including storage unit operating parameters, environmental parameters, personnel inspection records, etc., which have been denoised and standardized) is called; on the other hand, the model trained in S23 (including optimized feature extraction layer and prediction output layer parameters) is activated, ensuring that the data format fully matches the model input interface, eliminating the need for secondary processing.

[0100] The specific operation consists of two steps: The first step is feature extraction and vector set construction. Through the model's feature extraction layer (such as a long short-term memory network with an attention mechanism), three types of time-sensitive features are selectively extracted from the input time-series data: storage unit operating condition features (such as the fluctuation range of tank pressure within 1 hour, the interval between pump start-up and shutdown), environmental dynamic features (such as the 3-hour upward trend of ambient temperature, the influence coefficient of wind speed on gas diffusion), and human intervention features (such as the response time after an anomaly is discovered during inspection, the recovery trend of equipment parameters after maintenance). These features are then integrated according to the structure of "timestamp-feature type-feature value" to form a sensitive feature vector set (e.g., "10:00-operating condition feature-pressure fluctuation 0.2M"). Pa; 10:00 - Environmental dynamic characteristics - temperature rise of 2℃; 10:00 - Human intervention characteristics - no abnormalities found during inspection); The second step is the output of the risk level time series curve. The sensitive feature vector set is input into the prediction output layer (fully connected network structure). The model will calculate the risk level of each time node within a preset time period (which can be set as needed, such as 24 hours for daily monitoring and 48 hours for extreme weather) based on the "feature-risk" mapping relationship learned during the training phase. Finally, the time series curve is output. For example, the curve will show the dynamic changes of "12:00 - overall regional risk level I (low risk), unit C risk level II; 15:00 - overall regional risk level II, unit C risk level III (high risk)", while distinguishing between overall and individual unit risks. The output risk level time series curve will be directly used for subsequent S30 (marking high-risk time window) and S40 (determining target storage unit), becoming a key bridge connecting time dynamic prediction and spatial diffusion analysis, ensuring that subsequent steps can be carried out based on accurate time dimension risk data.

[0101] In one embodiment, step S24 involves inputting the sensitive feature vector set into the prediction output layer, and outputting a risk level time series curve for a preset future duration through the prediction output layer. This specifically includes the following steps:

[0102] S241. Associate the sensitive feature vector set with the hazardous chemical type attribute of the corresponding storage unit to form an associated dataset; based on the associated dataset, group the sensitive feature vector set according to the hazardous chemical type, and sort the sensitive feature vectors in each group according to the time series to form a time series feature matrix hierarchically based on the risk characteristics of hazardous chemicals.

[0103] S242. Analyze the dynamic correlation relationship between adjacent vectors in each hierarchical time-series feature matrix through a preset time-series correlation algorithm, and based on the attention mechanism, assign a preset high-impact weight value to the accident-related feature vectors that match the historical accident features, and assign a preset basic weight value to other non-accident-related auxiliary feature vectors; wherein, the historical accident features are extracted based on historical accident data of the same type of storage unit within a preset time period.

[0104] S243. Based on the preset hazardous chemical type risk level mapping rules, combined with the risk level trend correction factor, and integrating the high impact weight value of the accident-related feature vector with the basic weight value of the auxiliary feature vector, calculate the risk level value for each period within the preset time period in the future.

[0105] S244. Arrange the risk level values ​​of each time period continuously along the time axis to generate a risk level time series curve. The risk level time series curve includes storage units marked with spatial association unit labels of the storage units located within the preset risk threshold range.

[0106] In this embodiment, as described in step S241, the core objective is to bind the sensitive feature vector set with the hazardous chemical type attribute, achieving a precise correlation between "features and material properties." This avoids the masking of key characteristics due to mixed analysis of risk features of different types of hazardous chemicals (e.g., the risk evolution patterns of flammable liquids and corrosive solids differ significantly), providing structured input for the differentiated calculation of the subsequent prediction output layer. For the input data, on the one hand, the sensitive feature vector set extracted in S24 (containing three types of features: operating conditions, environment, and human intervention) is used; on the other hand, the hazardous chemical type attribute of each storage unit is called (e.g., "Unit A - Ethanol (flammable liquid)", "Unit B - Sodium hydroxide (corrosive solid)", "Unit C - Ammonia (toxic gas)") to ensure that the feature vector corresponds one-to-one with the specific hazardous chemical type.

[0107] The specific operation consists of three steps: The first step is to associate and form a dataset, binding each sensitive feature vector (e.g., the pressure fluctuation feature of unit A at 10:00) with the hazardous chemical type attribute of that unit (e.g., ethanol), forming an associated dataset of "feature vector - hazardous chemical type," for example: "Feature: [Pressure fluctuation 0.2MPa, temperature rise 2℃]; Type: ethanol"; The second step is to group by type, dividing the sensitive feature vector set into several subgroups based on the associated dataset. Within the same subgroup are feature vectors of the same type of hazardous chemical. For example, all feature vectors of the "ethanol" unit are grouped... The quantities are categorized into the "Flammable Liquids" group, and the "Sodium Hydroxide" unit is categorized into the "Corrosive Solids" group. The third step is time-series sorting and matrix construction. Within each group, the sensitive feature vectors are arranged in chronological order by timestamp (e.g., sorted by hour from 9:00 to 18:00) to form a structured time-series feature matrix. For example, in the "Flammable Liquids" matrix, the rows represent time nodes (9:00, 10:00, etc.), the columns represent feature types (pressure fluctuations, temperature trends, etc.), and the matrix elements are the corresponding feature values, clearly presenting the evolution of the characteristics of the same type of hazardous chemicals over time.

[0108] The output time-series feature matrix, stratified by the risk characteristics of hazardous chemicals, will be directly input into the subsequent calculation stage of the prediction output layer. This enables the model to perform differentiated analysis on the risk characteristics of different types of hazardous chemicals (such as flammable, corrosive, and toxic chemicals). For example, for flammable chemicals, temperature / concentration is more important, while for corrosive chemicals, equipment pressure / sealing is more important. This lays the foundation for accurately outputting the time-series curve of the risk level.

[0109] As described in step S242, the core objective is to enhance the model's ability to identify "risk precursor features." By quantifying the dynamic correlation between features and assigning differentiated weights, it avoids key accident features from being overwhelmed by a massive amount of conventional features (such as distinguishing between pressure surge features directly related to the accident and ordinary environmental temperature and humidity features), thus providing a more accurate basis for the importance of features in subsequent risk level calculations. For the input data, on the one hand, the "time-series feature matrix stratified by hazardous chemical type" (such as the feature matrix for flammable liquids and corrosive solids) output in S241 is used; on the other hand, "historical accident features of the same type of storage unit" (such as the "pressure rise of 0.5 MPa within 1 hour + gas concentration exceeding the standard" combination feature extracted from ethanol storage tank leakage accident data) is called to ensure that the weight allocation is supported by actual accident data.

[0110] The specific operation consists of two steps: The first step is dynamic correlation analysis, which uses preset time-series correlation algorithms (such as sliding window correlation analysis and time-series pattern mining algorithms) to capture the dependencies between adjacent time vectors in the hierarchical matrix. For example, in the flammable liquid group matrix, the correlation strength between "pressure fluctuation characteristics at time t" and "gas concentration characteristics at time t+1" is analyzed (such as the probability of concentration exceeding the standard within 10 minutes after a sudden pressure rise), to identify typical risk evolution chains such as "pressure anomaly → concentration exceeding the standard". The second step is weighting based on the attention mechanism, which compares the current feature vector with historical accident characteristics. The feature vectors are compared, and if a match is found (e.g., the current "pressure rise of 0.4 MPa in 1 hour in the ethanol unit" matches the historical accident feature "pressure rise ≥ 0.3 MPa"), it is identified as an "accident-related feature vector" and assigned a high-impact weight value (e.g., weight value 0.8). Unmatched auxiliary feature vectors (e.g., normal fluctuations in ambient humidity) are assigned a basic weight value (e.g., weight value 0.2). For example, in the feature matrix of the ethanol unit, the "pressure fluctuation vector" is assigned a high weight because it matches the historical leakage accident feature, while the "no abnormalities found in daily inspections vector" is assigned a basic weight as an auxiliary feature. The output "weighted hierarchical time-series feature matrix" clearly marks the impact weight of each feature vector, preserving the dynamic correlation information between features and highlighting key accident-related features through weight differences. This matrix will be directly used in subsequent risk level calculations, ensuring that the model prioritizes high-weighted risk precursor features during prediction, thus improving the warning sensitivity of the risk level time-series curve.

[0111] As described in step S243, the core objective is to transform the weighted feature vectors into quantifiable risk level values, thereby realizing the transition from feature analysis to risk assessment. This avoids distortions in risk level calculations due to the calculation logic being detached from the characteristics of hazardous chemicals or ignoring risk evolution trends (such as using a uniform standard to measure the risk of flammable and corrosive hazardous chemicals). This provides accurate quantitative data support for the subsequent output of risk level time series curves. For the input data, on the one hand, the "weighted hierarchical time series feature matrix" output in S242 (containing high-impact weights for accident-related features and basic weights for auxiliary features) is called. On the other hand, two types of key parameters are introduced: a preset hazardous chemical type risk level mapping rule (a "feature-risk" correspondence standard customized according to the characteristics of hazardous chemicals) and a risk level trend correction factor (a dynamic adjustment parameter reflecting the change of risk over time), ensuring that the calculation logic conforms to the actual risk evolution pattern.

[0112] The specific operation consists of three steps: The first step is to clarify the mapping rules. The preset mapping rules are customized for the risk characteristics of different types of hazardous chemicals. For example, the mapping rule for "ethanol (flammable liquid)" is "flammable gas concentration ≥15%LEL (high impact weight 0.8) → basic risk score 3 points; ambient temperature ≥35℃ (basic weight 0.2) → basic risk score 1 point"; the mapping rule for "sodium hydroxide (corrosive solid)" is "storage tank sealing leakage (high impact weight 0.8) → basic risk score 3 points; ambient humidity ≥60% (basic weight 0.2) → basic risk score 1 point", avoiding confusion of risk standards across different types of hazardous chemicals. The second step is to calculate the preliminary risk value. The feature vector of each time period is weighted and summed according to the logic of "feature basic risk score × corresponding weight". For example, for the ethanol unit, in a certain time period, "gas concentration 18%LEL (basic score 3 × 0.8 = 2.4) + temperature 36℃ (basic score 1 × 0.2 = 0.2)", the preliminary risk value = 2.4 + 0.2 = 2.6 points. The third step is to introduce a trend correction factor. If the characteristics before and after the period show a "continuous deterioration trend" (e.g., the gas concentration rises from 12% LEL to 18% LEL for three consecutive periods), the correction factor is 1.2 (risk is increased). If it shows a "stable trend", the correction factor is 1.0. The final risk level value = preliminary risk value × correction factor (e.g., 2.6 × 1.2 = 3.12 points, corresponding to Level III risk according to the preset classification standard).

[0113] The output "risk level value for each time period within the future preset time period" (e.g., risk level values ​​for 24 time periods in the next 24 hours at 1-hour intervals) will be directly used in subsequent steps of S24; these values ​​will be integrated according to the time series to form a complete risk level time series curve, ensuring that the risk level at each time node in the curve has a dual calculation basis of "characteristic adaptation + trend dynamics", thereby improving the accuracy and reliability of risk prediction.

[0114] As described in step S244, the core objective is to transform the dispersed time-period risk level values ​​into an intuitive and continuous time-series curve. By labeling key units, the curve simultaneously carries information on "temporal risk changes" and "spatial associated risks," avoiding the omission of spatially related hidden dangers (such as sensitive units surrounding high-risk units) in subsequent analyses due to only presenting single-time dimension data. This provides a visual carrier for "spatiotemporal information fusion" for the final output of S20, the labeling of high-risk time periods in S30, and the determination of target units in S40. For the input data, on the one hand, the "risk level value for each time period within the future preset duration" output by S243 is called (such as 24 risk level quantification values ​​at 24-hour and 1-hour intervals in the future). On the other hand, two types of key information are associated: the basic information of the storage unit after preprocessing in S10 (unit ID, type of stored hazardous chemicals), and the correspondence between "storage unit-spatial associated units" in the spatial topology data of S10 (such as adjacent units B and C of unit A, or nearby fire dikes, ventilation ducts, and other associated facilities). At the same time, the preset risk threshold is used (such as risk level ≥ III as the high-risk judgment standard).

[0115] The specific operation consists of two steps: The first step is curve generation, which involves arranging the risk level values ​​for each time period sequentially along a time axis (e.g., from 0:00 to 24:00) to construct a two-dimensional risk level time series curve: the horizontal axis represents time nodes (e.g., 0:00, 1:00...24:00), and the vertical axis represents risk levels (e.g., Level I to Level IV). A smooth curve connects the values ​​for each time period, clearly showing the dynamic trend of risk changes over time (e.g., the risk is stable at Level I from 0:00 to 8:00, rises to Level II at 9:00, and reaches Level III at 12:00). The second step is key unit labeling, which involves filtering out "time periods with excessive risk levels" based on preset risk thresholds. Two types of labels are marked on the curve positions at corresponding time nodes: one is "storage units located within the preset risk threshold range" (e.g., unit A, whose risk reaches level III at 12:00, is labeled "Unit A (ethanol) - Level III risk" and marked with a red triangle); the other is the "spatial related units" of this unit (e.g., adjacent units B and C of unit A are labeled "Related unit B (toluene) - Potential risk" and "Related unit C (fire hydrant) - Protective facilities" and marked with yellow dots). The labels simultaneously indicate the unit ID, stored substance, and risk association type (direct exceedance / potential impact). The final output risk level time series curve not only intuitively shows the risk evolution pattern in the time dimension, but also clarifies the exceedance units and their spatially related objects through labels, achieving a preliminary integration of "time risk - spatial association". This curve will directly serve as the core output result of step S20, providing key data support that is "visible and locatable" for the subsequent S30 to accurately locate high-risk periods and for S40 to determine the target storage units and related units that need to be analyzed in detail.

[0116] In one embodiment, the method for determining the preset risk threshold includes the following steps:

[0117] S201. Based on the hazardous chemical type attribute of the storage unit, retrieve the basic safety threshold of the corresponding type of hazardous chemical as the initial risk threshold.

[0118] S202. Extract the risk level threshold value of the corresponding type of hazardous chemical from the characteristics of historical accidents, and correct the initial risk threshold based on the risk level threshold value to obtain the type-specific benchmark threshold.

[0119] S203. Extract real-time dynamic impact factors from the set of sensitive feature vectors, dynamically adjust the type-specific benchmark threshold based on the extracted real-time dynamic impact factors, and determine the preset risk threshold of the corresponding storage unit.

[0120] In this embodiment, as described in steps S201-S203 above, the core objective is to construct a three-layer threshold determination logic of "type adaptation + historical verification + real-time dynamics" to avoid the problems of "high-risk misjudgment (threshold too high)" or "low-risk missed judgment (threshold too low)" caused by using fixed thresholds. This ensures that the risk level determination criteria in steps S20, S24, etc., not only conform to the characteristics of hazardous chemicals but also adapt to changes in actual scenarios. The three steps are sequentially linked, from basic thresholds to dynamic adjustments, gradually forming accurate preset risk thresholds, as detailed below:

[0121] Step S201 aims to establish initial threshold benchmarks that are "characteristically compatible" for different types of hazardous chemicals, avoiding the use of a uniform standard to measure the risks of hazardous chemicals with significant differences (such as the vastly different safety thresholds between flammable gases and corrosive liquids). For input data, the hazardous chemical type attribute of the storage unit is used (e.g., "Unit A - Ethanol (Flammable Liquid)", "Unit B - Ammonia (Toxic Gas)", which can be retrieved from the spatial topology data preprocessed in S10). During operation, the corresponding basic safety threshold is retrieved from the preset "Hazardous Chemical Safety Threshold Library" based on the type attribute. This threshold library is constructed based on national / industry standards; for example, the basic safety thresholds for ethanol are "Flammable gas concentration ≥ 20% LEL (Lower Explosive Limit)" and "Temperature ≥ 40℃", while the basic safety thresholds for ammonia are "Gas concentration ≥ 30ppm" and "Pressure ≥ 0.6MPa". The output initial risk thresholds provide a "type-specific" starting benchmark for subsequent corrections.

[0122] For step S202, the initial threshold is validated and optimized using historical accident data to ensure that the threshold closely matches the actual accident evolution pattern (avoiding a disconnect between the standard threshold and the characteristics of on-site accidents). For input data, on one hand, the initial risk threshold output from S201 is called; on the other hand, the risk level threshold value (i.e., the actual risk parameter at the time of the accident) is extracted from the "historical accident characteristics of similar storage units" in S21. This is because, in historical ethanol leak accidents, the actual pre-explosion gas concentration threshold value is 15% LEL, lower than the initial threshold of 20% LEL. During operation, the threshold value is compared with the initial threshold. If the threshold value is lower than the initial threshold (indicating an earlier actual risk trigger point), the initial threshold is lowered to near the threshold value (e.g., the ethanol concentration threshold is corrected from 20% LEL to 16% LEL). If the threshold value is higher than the initial threshold, the initial threshold is maintained or slightly adjusted. The output type-specific benchmark threshold, which integrates standard requirements and historical accident experience, is more applicable to the field than the initial threshold.

[0123] For step S203, real-time factors are introduced to dynamically calibrate the baseline threshold, avoiding the inability to cope with sudden scenarios (such as extreme weather, equipment aging, and other temporary risk enhancement factors) due to fixed thresholds. For input data, on the one hand, the type-specific baseline threshold output from S202 is called, and on the other hand, real-time dynamic influencing factors (such as sudden increase in ambient temperature and humidity, equipment exceeding its service life, and the presence of nearby fire sources) are extracted from the sensitive feature vector set extracted in S24. During operation, adjustment rules are set according to the degree of influence of factors. For example, in high-temperature weather (ambient temperature ≥35℃), the ethanol concentration threshold is reduced by 20% from the baseline value of 16%LEL (to 12.8%LEL); when equipment is aging (operating for more than 10 years), the pressure threshold is reduced by 15%. The final output preset risk threshold retains the basis of type specificity and historical verification, and incorporates real-time scenario variables. It can be directly used as the judgment standard for "whether the risk level exceeds the standard" in S20 and S24, ensuring the accuracy and timeliness of risk identification.

[0124] In one embodiment, step S40 specifically includes the following steps:

[0125] S41. Perform hierarchical analysis on the preprocessed spatial topology data, extract the spatial association parameters and physical barrier levels between the target storage unit and adjacent storage units, and construct a spatial diffusion risk assessment model based on the diffusion characteristic parameters of hazardous chemicals within the target storage unit.

[0126] S42. When the risk level time series curve is associated with a storage unit within the preset risk threshold range, the real-time risk level value of the target storage unit is obtained and input into the spatial diffusion risk assessment model. The risk diffusion characteristics of the target storage unit are extracted through the spatial diffusion risk assessment model. The risk diffusion characteristics include the risk diffusion rate, diffusion radius and hazardous substance concentration decay coefficient of the target storage unit.

[0127] Based on the extracted risk diffusion characteristics and the spatial correlation parameters between the target storage unit and adjacent storage units, S43 associates the risk diffusion characteristics of the target storage unit with the corresponding adjacent storage units according to the preset correlation rules, calculates the risk level of each adjacent storage unit affected by diffusion, and generates a correlation table containing the target storage unit, adjacent storage units, diffusion path and affected risk level.

[0128] S44. Based on the risk change trend of the risk level time series curve, the association table is overlaid and matched with the spatial map of the storage area to generate and output the dynamic risk spatial distribution result of the storage area that is dynamically updated over time. The dynamic risk spatial distribution result includes the actual risk level and diffusion impact range of each storage unit at different time periods.

[0129] In this embodiment, as described in step S41, the core objective is to build a model framework that can accurately simulate the diffusion path and impact range of hazardous chemicals. This provides a computational basis for subsequent analysis of the transmission pattern of risk from the target unit to the surrounding area, avoiding distortion in the diffusion range prediction (e.g., underestimating the diffusion rate of flammable gases or overestimating the impact radius of liquid leaks) due to the model ignoring spatial relationship details (such as unit spacing and barrier facilities) or the diffusion characteristics of hazardous chemicals (such as density and volatilization rate). For the input data, on the one hand, the spatial topology data preprocessed in S10 (including the location, spacing, and connectivity of pipes / fire dikes, etc.) is used; on the other hand, the diffusion characteristic parameters of hazardous chemicals within the target storage unit (such as the volatilization rate of ethanol, the diffusion coefficient of ammonia, and the flow viscosity of sulfuric acid, which can be retrieved from the hazardous chemical safety technical specifications) are called to ensure that the model input simultaneously covers both spatial structure and material characteristics.

[0130] The specific operation consists of three steps: The first step is to analyze the spatial topology data in layers, breaking the data down into "unit layer - association layer - barrier layer"; the unit layer focuses on the physical attributes of the target storage unit and adjacent units (such as tank volume and material); the association layer extracts the spatial association parameters between units (such as the straight-line distance between the target unit and adjacent units, the pipe connection method (direct connection / indirect connection), and the relative height difference); the barrier layer determines the physical barrier level (such as the height and fire resistance rating of the firewall, the material of the wall (concrete / sheet metal), and the coverage of the ventilation facilities, which are divided into "strong barrier (level 3) - medium barrier (level 2) - weak barrier (level 1)" according to the protection capacity); the second step is parameter integration, which integrates the spatial associations. The first step involves associating parameters (such as a distance of 50 meters and direct pipeline connection), physical barrier levels (such as a firewall level 3), and diffusion characteristic parameters (such as an ethanol evaporation rate of 0.5 kg / h) to form a model input parameter set. The second step is model construction, which selects a suitable diffusion algorithm based on the parameter set (e.g., a "gravity flow + evaporation diffusion" model for flammable liquids and a "Gaussian diffusion model" for toxic gases) and embedding barrier attenuation coefficients (e.g., a level 3 barrier can reduce diffusion impact by 60%). For example, the model for an ethanol storage tank (target unit) calculates "the concentration of unit B 50 meters away from the barrier when there is no barrier, multiplied by the attenuation coefficient of the firewall (level 3) of 0.4, to obtain the actual affected concentration." The output spatial diffusion risk assessment model has the function of "inputting target unit information → calculating diffusion range, affected concentration, and the level of impact on adjacent units," and will be directly used in the subsequent step S40: "When the target storage unit is marked as high risk, this model quickly extracts its risk diffusion characteristics, providing a core calculation tool for outputting dynamic risk spatial distribution results."

[0131] As described in step S42, the core objective is to transform the high-risk units identified in the time dimension (S20 annotation results) into quantifiable spatial diffusion parameters. This provides key characteristic evidence for analyzing the transmission patterns of risk to the surrounding areas, avoiding the situation where we only know that "the unit has high risk" but do not understand "how the risk spreads and the extent of its impact," leading to a lack of targeted spatial control measures. For the input data, on the one hand, we call up the "high-risk storage units associated with the time-series risk level curve" (i.e., target storage units, such as unit A with a risk level ≥ III) and their real-time risk level values ​​(such as unit A's current risk level being IV) output by S20. On the other hand, we activate the spatial diffusion risk assessment model constructed in S41 (including spatial association parameters, physical barrier levels, and hazardous chemical diffusion characteristic parameters) to ensure that the input matches the format and dimensions required by the model.

[0132] The specific operation consists of two steps: The first step is triggering and data acquisition. When the system detects a storage unit within a preset risk threshold range in the risk level time series curve (e.g., unit A has a risk level of IV, exceeding the threshold of III), it automatically triggers the diffusion analysis process and simultaneously acquires the real-time risk level value of the target unit (reflecting the current severity of the risk), unit ID, and corresponding hazardous chemical type (e.g., unit A stores ethanol). The second step is model calculation and feature extraction. The above data is input into the spatial diffusion risk assessment model. The model calculates and extracts three core risk diffusion characteristics based on a preset algorithm (e.g., combining parameters such as the evaporation rate of ethanol and the current wind speed): risk diffusion rate (e.g., ethanol vapor diffuses to the surrounding area at a speed of 0.3 m / s), diffusion radius (e.g., the maximum diffusion range is 50 meters within 1 hour), and hazardous substance concentration decay coefficient (e.g., the ethanol concentration decreases by 30% for every 10 meters increase in distance from unit A). For example, the diffusion characteristic calculation result for unit A (ethanol, level IV risk) is "rate 0.3 m / s, radius 50 meters, decay coefficient 0.3 / 10 meters", clearly reflecting the propagation law of risk in space. The output risk diffusion characteristics (rate, range, attenuation coefficient) will be directly used in subsequent steps of S40, linked to adjacent storage units, to calculate the risk level of adjacent units, providing quantitative diffusion parameters to support the final output of "dynamic risk spatial distribution results of storage area", ensuring that spatial risk analysis knows both the "risk source" and the "diffusion path and degree of impact".

[0133] As described in step S43, the core objective is to establish a risk transmission correlation between the target unit and adjacent units, quantify the degree to which adjacent units are affected by diffusion, and avoid focusing only on the risk of the target unit itself while ignoring the surrounding chain risks (such as adjacent units being rapidly affected due to proximity or lack of barriers after the target unit leaks). This provides structured correlation data support for the subsequent generation of dynamic risk spatial distribution results. For the input data, on the one hand, the risk diffusion features extracted in S42 (diffusion rate, radius, and concentration decay coefficient of the target unit) are called, and on the other hand, the spatial correlation parameters analyzed in S41 (distance between the target unit and adjacent units, relative position, pipeline connection relationship, physical barrier level, etc.) are used. At the same time, preset correlation rules are introduced (defining the logic of "which adjacent units should be correlated" and "how the risk level is calculated") to ensure that the correlation and calculation logic is quantifiable and reproducible.

[0134] The specific operation consists of three steps: The first step is to define the associated range. Based on the preset association rules, adjacent units that are "within the diffusion radius of the target unit" and whose "spatial association parameters meet the conduction conditions" are selected. For example, if the diffusion radius of target unit A (ethanol) is 50 meters, and the rule is set to "units with a spacing ≤ 50 meters and no insurmountable barriers (such as concrete walls over 3 meters) are associated objects", then unit B (toluene, separated by only a 1-meter high-speed iron wall) at 30 meters and unit C (empty tank area, no barrier) at 45 meters are selected, while unit D (outside the range) at 60 meters is excluded. The second step is to calculate the affected risk level. Combining diffusion characteristics and spatial parameters, the risk level is calculated according to "concentration decay × barrier". The logical calculation of "attenuation": Unit B is 30 meters away, and the concentration attenuation coefficient is "30% attenuation every 10 meters" (90% attenuation over 30 meters, leaving 10%). The physical barrier (iron wall) is level 2 (attenuation of 40%). Therefore, the actual affected concentration = initial concentration × 10% × (1-40%) = initial concentration × 6%. According to the ethanol diffusion risk level standard (concentration ≥ 5% corresponds to level II risk), the affected level of Unit B is level II. Unit C is 45 meters away, without a barrier, and the concentration attenuates to 15% × 100% = 15%, corresponding to level III risk. The third step is to generate a relational table, which records the data in a structured manner according to "target unit - adjacent unit - diffusion path - affected level", for example:

[0135] Table 1: Relationship Table

[0136]

[0137] The output correlation table clearly presents the specific chain and degree of impact of the risk transmission from the target unit to the surrounding area. It will be directly used in the subsequent steps of S40: to integrate it into the dynamic risk spatial distribution results of the storage area, providing core data for the visualization of "risk source-diffusion path-affected area", and ensuring that the spatial risk analysis is both correlated and quantitative.

[0138] As described in step S44, the core objective is to deeply integrate the "relationship of risk diffusion" with "spatial geographic information" and "temporal change trends" to form an intuitive dynamic spatial risk picture. This avoids the problems of "vague spatial location" and "lack of temporal dynamics" caused by describing risk distribution only through tables or text (such as the inability to intuitively judge the movement trajectory of high-risk areas at different times). This provides "visible and traceable" spatial dimension dynamic data support for the spatiotemporal coupling analysis of S50. For the input data, on the one hand, the correlation table output by S43 (including target unit, adjacent units, diffusion path and affected level) is called. On the other hand, two types of key information are introduced: the risk level time series curve output by S20 (reflecting the trend of risk change over time, such as "10:00-12:00" the risk of the target unit continues to increase) and the spatial map of the storage area (generated based on the spatial topology data preprocessed by S10, including geographic information such as unit location coordinates, roads, and protective facilities), ensuring that the input data simultaneously covers the three dimensions of "relationship", "temporal trend" and "spatial geographic".

[0139] The specific operation consists of three steps: The first step is dynamic matching along the time dimension. Based on the time nodes of the risk level time series curve (such as 10:00, 11:00, and 12:00 divided into 1-hour intervals), the risk data in the correlation table is split according to the corresponding time period. For example, "the diffusion characteristics of unit A at 10:00" is bound to "the correlation relationship at 10:00", and "the diffusion characteristics at 11:00" is bound to "the correlation relationship at 11:00", ensuring that the risk data at each time node corresponds independently. The second step is spatial map overlay. The correlation data of each time period (target unit location, adjacent unit location, diffusion path, and affected level) is mapped to the corresponding coordinates on the spatial map, and the risk level is distinguished by visual symbols (such as red). The risk levels are marked with color (Level III and above, yellow for Level II), diffusion range (e.g., a dashed box indicates a 50-meter diffusion range at 10:00, and a solid box indicates a 60-meter diffusion range at 11:00), and diffusion path (e.g., arrows indicating the diffusion direction from "Unit A to Unit B"). The third step is dynamic updating and integration, which connects the spatial overlay results of each time period in chronological order to form a dynamic view that automatically updates over time. For example, at 10:00, the map displays "Unit A (red), Unit B (yellow), diffusion range 50 meters"; at 11:00, due to the risk escalation, it displays "Unit A (dark red), Unit B (orange), newly added Unit C (yellow), diffusion range 60 meters," intuitively presenting the expansion trend of the risk in space.

[0140] The output of dynamic risk spatial distribution results is presented in the form of dynamic maps or time-series slices, and its core includes three elements: "time axis + spatial map + risk labeling". The time axis can be dragged to select any time period, and the spatial map displays in real time the actual risk level (including its own risk and the risk of being affected by diffusion) of each storage unit during that time period, the boundary of the diffusion impact range, and key diffusion paths. This result will be directly used for the spatiotemporal coupling analysis of S50, accurately matching it with the high-risk time period window (S30), providing a visual basis for identifying the overlapping area of ​​"high-risk time period + high-risk spatial area", and ensuring that subsequent risk warning and control measures can be accurately located to the time and place.

[0141] In one embodiment, step S42 further includes the following step:

[0142] S421. Obtain the physical properties of the target storage unit and the hazardous chemical type properties inside it, and determine whether the target storage unit belongs to a special scenario based on a preset judgment threshold.

[0143] S422. When the target storage unit belongs to a special scenario, a preset special scenario correction coefficient table is called to match the correction coefficient corresponding to the target storage unit. The special scenario correction coefficient table contains correction coefficients corresponding to several special scenarios.

[0144] S423. Dynamically correct the extracted risk diffusion characteristics according to the correction coefficient corresponding to the target storage unit to obtain the corrected risk diffusion characteristics.

[0145] In this embodiment, steps S421-S423 are supplementary optimization steps to extract the risk diffusion characteristics of the target storage unit in step S42. The core purpose is to correct the diffusion characteristics for special scenarios (such as storage tanks with heating devices or units storing highly toxic substances), avoiding distortion in the calculation of parameters such as diffusion rate and range due to ignoring the risk amplification / attenuation effects of special scenarios (e.g., heated storage tanks accelerate the evaporation of flammable liquids, so the diffusion rate should be higher than that of ordinary storage tanks), ensuring that the risk diffusion characteristics are more consistent with the actual scenario. The three steps sequentially realize a logical closed loop of determining special scenarios → matching correction coefficients → correcting diffusion characteristics, as detailed below:

[0146] Step S421 aims to accurately identify storage units requiring special handling, providing a basis for subsequent corrections and avoiding the confusion between ordinary and special scenarios in calculations. For the input data, on the one hand, it retrieves the physical properties of the target storage unit (such as whether the tank material is insulated, whether it is equipped with heating / cooling devices, and whether the tank volume exceeds 500). On the one hand, it checks whether the location is underground or semi-underground; on the other hand, it retrieves the type and attributes of hazardous chemicals within the unit (such as whether they are highly toxic substances, spontaneously combustible substances, or substances that react with water, which can be obtained from the hazardous chemical safety data sheet); at the same time, it introduces a preset judgment threshold, which is based on the "special scenario judgment standard" defined by industry safety standards, such as: "storage tanks equipped with heating devices", "units storing highly toxic substances (LD50≤50mg / kg)", "underground storage tanks with a volume exceeding 1000 cubic meters". "All of these were determined to be special scenarios."

[0147] In practice, the physical properties and hazardous chemical type attributes of the target unit are compared one by one with preset judgment thresholds: if any threshold condition is met, it is judged as a special scenario; if none are met, it is a normal scenario. For example, target unit E is a "methanol storage tank equipped with a steam heating device (methanol is a flammable liquid, not highly toxic)," and its physical property "equipped with a heating device" meets the preset threshold, therefore it is judged as a special scenario; target unit F is an "ethanol storage tank made of ordinary carbon steel (without a heating device, volume 300..."). If the condition is "Yes / No Special Scenario" and the corresponding special scenario type (such as "Heating Tank Scenario" or "Highly Toxic Substance Scenario"), then it is determined to be a normal scenario. The output result is "Yes / No Special Scenario" and the corresponding special scenario type (such as "Heating Tank Scenario" or "Highly Toxic Substance Scenario").

[0148] Step S422 aims to provide quantitative correction criteria for different special scenarios, avoiding subjective correction logic (such as judging the impact of heating devices based on experience). For the input data, on the one hand, the "special scenario type" output from S421 (such as "heated storage tank scenario" or "highly toxic substance scenario") is used; on the other hand, a preset special scenario correction coefficient table is called. This table is constructed based on historical experimental data and accident cases, divided according to "scenario type - diffusion characteristic dimension," clearly defining the correction ratio for "diffusion rate, diffusion radius, and concentration decay coefficient" for different scenarios. An example is shown below:

[0149] Table 2: Correction Coefficients for Special Scenarios

[0150]

[0151] In practice, based on the scenario type determined by S421, the corresponding correction coefficient is precisely matched in the table. For example, if S421 determines unit E as "heated storage tank scenario (temperature 55℃)," then "diffusion rate coefficient 1.2, range coefficient 1.1, attenuation coefficient 0.9" is matched; if it is determined as "highly toxic substance scenario," then "rate coefficient 1.0, range coefficient 1.3, attenuation coefficient 0.8" is matched. The output results are the three sets of correction coefficients corresponding to this specific scenario (corresponding to diffusion rate, range radius, and concentration attenuation coefficient, respectively).

[0152] Step S423 is the implementation step for optimization in specific scenarios. It adjusts the original diffusion characteristics by quantifying coefficients to ensure that the results conform to the actual risk patterns of the specific scenario. For the input data, on the one hand, the three sets of correction coefficients output by S422 are called, and on the other hand, the "initial risk diffusion characteristics" extracted in the original step S42 (such as the diffusion rate before correction of 0.3m / s, the range radius of 50m, and the concentration decay coefficient of 0.3 / 10m) are used.

[0153] In practice, the corrections are made one by one according to the logic of "initial eigenvalue × corresponding correction coefficient":

[0154] Corrected diffusion rate = initial diffusion rate × diffusion rate correction factor (e.g., 0.3 m / s × 1.2 = 0.36 m / s).

[0155] The corrected diffusion radius = initial radius × radius correction factor (e.g., 50m × 1.1 = 55m).

[0156] Corrected concentration decay coefficient = initial decay coefficient × concentration decay coefficient correction coefficient (e.g., 0.3 / 10m × 0.9 = 0.27 / 10m).

[0157] For example, the initial diffusion characteristics of unit E are "velocity 0.3 m / s, range 50 m, attenuation 0.3 / 10 m". After correction by the heating scenario coefficient, it becomes "velocity 0.36 m / s, range 55 m, attenuation 0.27 / 10 m", which better reflects the actual situation where methanol volatilizes faster, the diffusion range expands, and the concentration attenuation slows down under heating conditions. The output result is "corrected risk diffusion characteristics" (including corrected velocity, range, and attenuation coefficient), which will directly replace the original initial characteristics and be used in step S43 to calculate the risk level of adjacent units, ensuring more accurate spatial diffusion analysis under special scenarios.

[0158] In one embodiment, step S50 specifically includes the following steps:

[0159] S51. Extract the time boundary parameters of the high-risk period window, wherein the time boundary parameters include the high-risk start time, duration and the peak time of the risk level;

[0160] S52. Map the time boundary parameters of the high-risk period window to the dynamic risk spatial distribution results in a spatiotemporal dimension, match the risk spatial distribution status of the storage area corresponding to each time node, and form an initial spatiotemporal coupling matrix. The initial spatiotemporal coupling matrix contains the associated data of time nodes, storage units, risk levels, and diffusion range.

[0161] S53. The initial spatiotemporal coupling matrix is ​​modified based on preset risk association rules, which include time priority weighting rules, spatial correlation rules and risk superposition rules.

[0162] S54. Generate and output spatiotemporal coupling risk warning information based on the corrected spatiotemporal coupling matrix.

[0163] In this embodiment, S51-S54 are the core sub-steps in S50 that realize spatiotemporal coupling analysis and early warning output. The core purpose is to accurately correlate high-risk periods in the time dimension with risk distribution in the spatial dimension through a closed-loop process of "time parameter extraction → spatiotemporal mapping → rule correction → early warning generation," ultimately outputting risk early warning information with both spatiotemporal accuracy. This avoids "misalignment of control timing" or "regional positioning deviation" caused by the fragmentation of spatiotemporal analysis (such as knowing only the high-risk period but not the corresponding high-risk area). The four steps are progressively advanced, gradually realizing the transformation from "data association" to "decision support," as detailed below:

[0164] Step S51 aims to clarify the key time nodes of high-risk periods, providing accurate time coordinates for subsequent spatiotemporal mapping and avoiding mapping deviations caused by ambiguous time boundaries (such as the inability to determine the specific high-risk period corresponding to a certain spatial state). For input data, the "high-risk period window" output by S30 is used (e.g., "8:00-10:30 (high risk in unit A)", "14:00-16:00 (high risk in unit A)"). During operation, three types of core time boundary parameters are extracted from each window: the high-risk start time (e.g., 8:00, 14:00, marking the time point when the risk begins), the duration (e.g., 2.5 hours, 2 hours, reflecting the time span of the risk's duration), and the peak time of the risk level (e.g., 9:15, 15:00, the highest risk time point determined based on the risk level time series curve of S20).

[0165] For example, for the "8:00-10:30" window, the extracted parameters are "starting at 8:00, lasting 2.5 hours, peak at 9:15". The output time boundary parameters will serve as the reference coordinates for the "time dimension" in S52, ensuring that each time node accurately corresponds to the spatial state.

[0166] Step S52 aims to establish a one-to-one correspondence between time nodes and spatial states, integrating fragmented time and spatial data into a structured matrix to avoid the loss of correlation information caused by analyzing the spatiotemporal dimensions separately (e.g., not knowing the high-risk spatial distribution corresponding to the peak time of 9:15). For the input data, on the one hand, the time boundary parameters extracted in S51 (key nodes such as start time and peak time) are called, and on the other hand, the "dynamic risk spatial distribution results" output by S44 (including the unit risk level and diffusion range of different time periods) are used. During operation, the mapping is performed according to the logic of "time node → corresponding spatial state": for the start time (e.g., 8:00), the spatial distribution at that time is matched (e.g., unit A is at risk level III, diffusion range is 40 meters, and unit B is at risk level II); for the peak time (e.g., 9:15), the spatial distribution at that time is matched (e.g., unit A is at risk level IV, diffusion range is 50 meters, and units B and C are at risk level III). Finally, it is integrated into an initial spatiotemporal coupling matrix, as shown in the example below:

[0167] Table 3: Initial Spatiotemporal Coupling Matrix

[0168]

[0169] The output initial spatiotemporal coupling matrix clearly presents the risk status and diffusion range of each storage unit at different time points, providing a basic data framework for subsequent corrections.

[0170] Step S53 aims to optimize the matrix through rules, highlighting high-priority risks (such as the overlapping risks of peak times and sensitive areas) and avoiding interference from low-priority information in decision-making (such as the slight diffusion risk during off-peak periods). For the input data, the initial spatiotemporal coupling matrix output from S52 is used, and the "preset risk association rules" (three core rules) are invoked. During operation, the matrix is ​​modified according to these rules:

[0171] Time priority weighting rules: assign 1.2 times the weight to the risk level at peak times (e.g., Unit B at 9:15 is affected by Level III → revised to Level III+), and assign 0.8 times the weight to off-peak times (e.g., 8:30) (Level II → Level II-), highlighting the time point with the most severe risk;

[0172] Spatial correlation rule: If the affected unit is close to a sensitive target (such as a water source or residential area), the risk level is raised by 1 level (e.g., if unit C is affected by Level III at 9:15 and is close to a water source → it is revised to Level IV).

[0173] Risk stacking rule: If the same unit is continuously affected during multiple high-risk periods (e.g., Unit B is at Level II or above during 8:00-10:30 and 14:00-16:00), it is marked as a "key focus unit".

[0174] For example, after correction, the risk level of Unit C (near the water source) at 9:15 AM was upgraded from Level III to Level IV, while Unit B, due to its continued impact, was marked as "highly concerned." The output corrected spatiotemporal coupling matrix eliminates redundant information, focuses on high-priority risks, and provides accurate data for early warning generation.

[0175] Step S54 aims to transform the corrected matrix into intuitive and actionable early warning information, realizing the transformation from "data matrix" to "decision-making basis" and avoiding the inability of on-site personnel to quickly understand risks due to the complexity of the data format. For input data, the spatiotemporal coupling matrix corrected in S53 is used; during operation, information is integrated according to the structure of "spatiotemporal overlay area → early warning level → impact indication":

[0176] Identify high-risk areas with overlapping time and space (e.g., "9:15-9:45 (peak period) + within 50 meters of the diffusion range of Unit A (including Unit C and the water source to the north)").

[0177] Early warning levels are classified according to the degree of risk overlap (e.g., "Red Alert - Emergency Response" and "Orange Alert - Key Prevention and Control").

[0178] Additional risk impact warnings (such as "Red alert areas may affect surrounding residential areas within 30 minutes, requiring immediate isolation").

[0179] The spatiotemporal coupled risk warning information output can be presented in the form of visual maps (marking high-risk areas and time periods) and text instructions (clearly specifying handling requirements), and can be directly used as the core input for S60 to generate control decision instructions, ensuring that control measures can accurately match the "high-risk spatiotemporal overlap area" and improve the efficiency of risk handling.

[0180] In one embodiment, a hazardous chemical storage accident risk analysis system is provided, which corresponds to the hazardous chemical storage accident risk analysis method described in the above embodiments. This hazardous chemical storage accident risk analysis system includes:

[0181] The data acquisition module is used to collect and preprocess multi-source data on hazardous chemical storage in the storage area. The multi-source data on hazardous chemical storage includes time-series data and spatial topology data.

[0182] The risk prediction module is used to input pre-processed time series data into a pre-built time dynamic risk prediction model, extract time-sensitive features from the time series data through the time dynamic risk prediction model, output a risk level time series curve for a future preset duration based on the time-sensitive features, and associate and label the storage units in the risk level time series curve that are within the preset risk threshold range.

[0183] The annotation module is used to annotate high-risk time period windows in the risk level time series curve where the risk level exceeds a preset safety threshold;

[0184] The risk assessment module is used to construct a spatial diffusion risk assessment model based on preprocessed spatial topology data. When the risk level time series curve is associated with a storage unit within the preset risk threshold range, the storage unit is taken as the target storage unit. The risk diffusion characteristics of the target storage unit are extracted through the spatial diffusion risk assessment model and associated with adjacent storage units, and the dynamic risk spatial distribution results of the storage area are output.

[0185] The coupling analysis module is used to perform spatiotemporal coupling analysis on high-risk time windows and dynamic risk spatial distribution results, correct the analysis results based on preset risk association rules, and output spatiotemporal coupling risk warning information.

[0186] The decision generation module is used to generate risk control decision instructions for hazardous chemical storage areas by calling a preset risk control rule base based on the spatiotemporal coupled risk warning information.

[0187] Specific limitations regarding a hazardous chemical storage accident risk analysis system can be found in the above description of the hazardous chemical storage accident risk analysis method, and will not be repeated here. Each module in the aforementioned hazardous chemical storage accident risk analysis system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0188] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows. Figure 2 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides the environment for the operation of the operating system and computer programs in the non-volatile storage media. The database is used for data storage, data processing, predictive analysis, etc. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for risk analysis of hazardous chemical storage accidents.

[0189] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements a method for risk analysis of hazardous chemical storage accidents.

[0190] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements a method for risk analysis of hazardous chemical storage accidents.

[0191] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0192] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0193] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural transformations made using the contents of the present invention's specification and drawings under the inventive concept of the present invention, or direct / indirect applications in other related technical fields, are included within the patent protection scope of the present invention.

Claims

1. A method for risk analysis of hazardous chemical storage accidents, characterized in that, Includes the following steps: Collect and preprocess multi-source data on hazardous chemical storage in the storage area, including time-series data and spatial topology data. The preprocessed time-series data is input into a pre-constructed time-dynamic risk prediction model. The model extracts time-sensitive features from the time-series data and outputs a time-series curve of risk level for a preset duration based on the time-sensitive features. It also associates and marks storage units in the time-series curve of risk level that are within a preset risk threshold range. The time-sensitive features include the hourly upward trend of temperature parameters, the frequency of abnormal fluctuations in gas concentration, and the duration of pressure exceeding the limit. The high-risk time period window in the risk level time series curve is marked to indicate that the risk level exceeds the preset safety threshold. Based on the preprocessed spatial topology data, a spatial diffusion risk assessment model is constructed. When the risk level time series curve is associated with a storage unit within the preset risk threshold range, the storage unit is taken as the target storage unit. The risk diffusion characteristics of the target storage unit are extracted through the spatial diffusion risk assessment model and associated with adjacent storage units, and the dynamic risk spatial distribution results of the storage area are output. The risk diffusion characteristics include the risk diffusion rate of the target storage unit, the diffusion radius, and the hazardous substance concentration decay coefficient. The high-risk time window is coupled with the dynamic risk spatial distribution results for spatiotemporal analysis. The analysis results are corrected based on the preset risk association rules, and spatiotemporal coupled risk warning information is output. Based on the spatiotemporal coupled risk warning information, a preset risk control rule base is invoked to generate risk control decision instructions for hazardous chemical storage areas.

2. The method for risk analysis of hazardous chemical storage accidents as described in claim 1, characterized in that, The construction of the time-dynamic risk prediction model specifically includes the following steps: Historical time-series sample data, including storage unit operating condition data, environmental dynamic data, and human intervention data, are acquired from the storage area. The historical time-series sample data is preprocessed to obtain a standardized time-series feature set. Based on historical accident data of the same type of storage unit within a preset time period, risk level labels are marked for the standardized time-series feature set. Data before the accident is marked as high risk, and normal operation data is marked as low risk. Construct a time-series prediction model that includes a feature extraction layer for extracting time-sensitive features and a prediction output layer for outputting time-series curves of risk levels; A standardized time-series feature set labeled with risk level is input into the time-series prediction model architecture, and the model parameters are adjusted using a preset training strategy to obtain a trained time-dynamic risk prediction model. The preprocessed time series data is input into the trained time dynamic risk prediction model. The time-sensitive features in the time series data are extracted through the feature extraction layer to form a sensitive feature vector set that includes storage unit operating condition features, environmental dynamic features, and human intervention features. The sensitive feature vector set is then input into the prediction output layer, which outputs the risk level time series curve for a future preset duration.

3. The method for risk analysis of hazardous chemical storage accidents as described in claim 2, characterized in that, The step of inputting the sensitive feature vector set into the prediction output layer and outputting a risk level time series curve for a preset future duration through the prediction output layer specifically includes the following steps: The sensitive feature vector set is associated with the hazardous chemical type attribute of the corresponding storage unit to form an associated dataset; based on the associated dataset, the sensitive feature vector set is grouped according to the hazardous chemical type, and the sensitive feature vectors in each group are sorted by time series to form a time series feature matrix hierarchically according to the risk characteristics of hazardous chemicals. The dynamic correlation between adjacent vectors in each hierarchical time-series feature matrix is ​​analyzed by a preset time-series correlation algorithm. Based on the attention mechanism, a preset high-impact weight value is assigned to the accident-related feature vectors that match the historical accident features, and a preset basic weight value is assigned to other non-accident-related auxiliary feature vectors. The historical accident features are extracted based on historical accident data of the same type of storage unit within a preset time period. Based on the preset risk level mapping rules for hazardous chemical types, combined with the risk level trend correction factor, and integrating the high impact weight value of the accident-related feature vector with the basic weight value of the auxiliary feature vector, the risk level value for each period within the preset time period is calculated. The risk level values ​​for each time period are arranged continuously along the time axis to generate a risk level time series curve. The risk level time series curve includes storage units marked with spatial association unit labels that are within the preset risk threshold range.

4. The method for risk analysis of hazardous chemical storage accidents as described in claim 3, characterized in that, The method for determining the preset risk threshold includes the following steps: Based on the hazardous chemical type attribute of the storage unit, the basic safety threshold of the corresponding type of hazardous chemical is retrieved as the initial risk threshold; The risk level threshold of the corresponding type of hazardous chemicals at the time of the historical accident is extracted from the characteristics of the historical accident. The initial risk threshold is corrected based on the risk level threshold to obtain the type-specific benchmark threshold. Real-time dynamic impact factors are extracted from the set of sensitive feature vectors. Based on the extracted real-time dynamic impact factors, the type-specific benchmark threshold is dynamically adjusted to determine the preset risk threshold of the corresponding storage unit.

5. The method for risk analysis of hazardous chemical storage accidents as described in claim 1, characterized in that, The step of constructing a spatial diffusion risk assessment model based on preprocessed spatial topology data, where a storage unit within a preset risk threshold range is associated with a time-series risk level curve, is used as the target storage unit. The risk diffusion characteristics of the target storage unit are extracted through the spatial diffusion risk assessment model and associated with adjacent storage units, outputting the dynamic spatial distribution result of the risk in the storage area. Specifically, the steps include the following: The preprocessed spatial topology data is analyzed hierarchically to extract spatial correlation parameters and physical barrier levels between the target storage unit and adjacent storage units. Based on the diffusion characteristic parameters of hazardous chemicals within the target storage unit, a spatial diffusion risk assessment model is constructed. When the risk level time series curve is associated with a storage unit within the preset risk threshold range, the real-time risk level value of the target storage unit is obtained and input into the spatial diffusion risk assessment model. The risk diffusion characteristics of the target storage unit are then extracted through the spatial diffusion risk assessment model. Based on the extracted risk diffusion characteristics and the spatial correlation parameters between the target storage unit and adjacent storage units, the risk diffusion characteristics of the target storage unit are associated with the corresponding adjacent storage units according to the preset correlation rules. The risk level of each adjacent storage unit affected by diffusion is calculated, and a correlation table containing the target storage unit, adjacent storage units, diffusion path and affected risk level is generated. Based on the risk change trend of the risk level time series curve, the correlation table is overlaid and matched with the spatial map of the storage area to generate and output the dynamic risk spatial distribution result of the storage area that is dynamically updated over time. The dynamic risk spatial distribution result includes the actual risk level and diffusion impact range of each storage unit at different time periods.

6. The method for risk analysis of hazardous chemical storage accidents as described in claim 5, characterized in that, The step of extracting the risk diffusion characteristics of the target storage unit through the spatial diffusion risk assessment model further includes the following steps: Obtain the physical properties of the target storage unit and the hazardous chemical type properties inside it, and determine whether the target storage unit belongs to a special scenario based on a preset judgment threshold; When the target storage unit belongs to a special scenario, a preset special scenario correction coefficient table is called to match the correction coefficient corresponding to the target storage unit. The special scenario correction coefficient table contains correction coefficients corresponding to several special scenarios. The extracted risk diffusion characteristics are dynamically corrected based on the correction coefficient corresponding to the target storage unit, resulting in the corrected risk diffusion characteristics.

7. The method for risk analysis of hazardous chemical storage accidents as described in claim 1, characterized in that, The step of performing spatiotemporal coupling analysis between the high-risk time period window and the dynamic risk spatial distribution results, correcting the analysis results based on preset risk association rules, and outputting spatiotemporally coupled risk early warning information specifically includes the following steps: Extract the time boundary parameters of the high-risk period window, including the high-risk start time, duration, and peak risk level time; The time boundary parameters of the high-risk period window are mapped to the dynamic risk spatial distribution results in a spatiotemporal dimension to match the risk spatial distribution status of the storage area corresponding to each time node, forming an initial spatiotemporal coupling matrix. The initial spatiotemporal coupling matrix contains the associated data of time node, storage unit, risk level, and diffusion range. The initial spatiotemporal coupling matrix is ​​modified based on preset risk association rules, which include time priority weighting rules, spatial correlation rules, and risk superposition rules. Based on the corrected spatiotemporal coupling matrix, spatiotemporal coupling risk warning information is generated and output.

8. A hazardous chemical storage accident risk analysis system, used to implement the steps of the hazardous chemical storage accident risk analysis method as described in any one of claims 1-7, characterized in that, include: The data acquisition module is used to collect and preprocess multi-source data on hazardous chemical storage in the storage area. The multi-source data on hazardous chemical storage includes time-series data and spatial topology data. The risk prediction module is used to input pre-processed time series data into a pre-built time dynamic risk prediction model, extract time-sensitive features from the time series data through the time dynamic risk prediction model, output a risk level time series curve for a future preset duration based on the time-sensitive features, and associate and label the storage units in the risk level time series curve that are within the preset risk threshold range. The annotation module is used to annotate high-risk time period windows in the risk level time series curve where the risk level exceeds a preset safety threshold; The risk assessment module is used to construct a spatial diffusion risk assessment model based on preprocessed spatial topology data. When the risk level time series curve is associated with a storage unit within the preset risk threshold range, the storage unit is taken as the target storage unit. The risk diffusion characteristics of the target storage unit are extracted through the spatial diffusion risk assessment model and associated with adjacent storage units, and the dynamic risk spatial distribution results of the storage area are output. The coupling analysis module is used to perform spatiotemporal coupling analysis on high-risk time windows and dynamic risk spatial distribution results, correct the analysis results based on preset risk association rules, and output spatiotemporal coupling risk warning information. The decision generation module is used to generate risk control decision instructions for hazardous chemical storage areas by calling a preset risk control rule base based on the spatiotemporal coupled risk warning information.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the hazardous chemical storage accident risk analysis method as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the hazardous chemical storage accident risk analysis method as described in any one of claims 1-7.

Images