A method and system for multi-reagent batch taking and permission linkage of a hazardous chemical cabinet

Abstract

The application discloses a hazardous chemical cabinet multi-reagent batch taking and permission linkage method and system, belonging to the field of computer technology, which realizes the quantitative evaluation of the distribution and evolution trend of the risk in the space and time dimensions by constructing a four-dimensional space-time risk field through real-time sensing of chemical incompatibility coupling risks in the multi-reagent batch taking scene and combining with the space position data in the cabinet. Therefore, the permission control strategy can be dynamically adjusted based on the real-time risk level, including the authentication factor combination mode, the reagent unlocking range and the monitoring sampling frequency. The method makes up for the problems of insufficient multi-reagent risk perception and rigid permission control of related technologies, and improves the safety management level of the hazardous chemical cabinet.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method and system for batch access and access control of multiple reagents in a hazardous chemical cabinet. Background Technology

[0002] Currently, intelligent management cabinets for hazardous chemicals generally adopt a control mode based on identity recognition and physical status monitoring. Related technical solutions typically use RFID readers, fingerprints, or facial recognition to identify the user, combined with weighing sensors, temperature and humidity sensors to record and provide early warnings about the storage status and environment of individual reagent bottles. When temperature or humidity exceeds limits or weight abnormalities are detected, an alarm signal is triggered to notify management personnel. Some high-end products implement remote authorized unlocking and dual-person, dual-lock security mechanisms to strengthen the control over the access to critical reagents.

[0003] However, the aforementioned technologies have significant limitations in practical applications. At the risk perception level, existing solutions only monitor the physical parameters of a single reagent bottle independently. When users take multiple reagents in bulk, they cannot detect the coupling risks that may arise from chemical incompatibility between different reagents, such as the potential for combustion and explosion when oxidizers and reducing agents are taken simultaneously. At the access control level, the authentication requirements and unlocking scope of these technologies are preset and fixed, disconnected from real-time risk status. When the risk increases, the system can only issue an alarm, unable to dynamically adjust permissions to prevent dangerous operations. Furthermore, the time dimension is only used as a timestamp for operation records and is not used for risk trend analysis and prediction.

[0004] Therefore, how to achieve real-time quantitative perception of the inherent risks of chemicals in scenarios involving the bulk use of multiple reagents, and how to establish a risk-driven dynamic permission linkage mechanism, has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] This application provides a method and system for batch access and access control of multiple reagents in a hazardous chemical cabinet, the technical solution of which is as follows: On the one hand, a method for batch access and access control of multiple reagents in a hazardous chemical cabinet is provided, the method comprising: In response to a bulk retrieval request containing at least two reagents to be retrieved, chemical kinetic parameters corresponding to the reagents to be retrieved are retrieved from a pre-stored chemical characteristic parameter library. The chemical kinetic parameters include reaction activation energy and reaction heat. The reaction activation energy is corrected by applying a temperature parameter from real-time environmental monitoring data to obtain a temperature-corrected activation energy parameter. Based on the temperature-corrected activation energy parameter, the heat of reaction, and the real-time environmental monitoring data, the coupling risk coefficients of the at least two reagents to be taken are calculated in real time for pairwise combinations, and the calculation results of all reagent combinations are aggregated to generate a multi-reagent coupling risk matrix. The multi-reagent coupling risk matrix is ​​used to quantitatively characterize the chemical incompatibility risk of all reagent combinations in the batch take-up list under real-time environment. The multi-reagent coupled risk matrix and the cabinet spatial location data obtained from the sensor network are input into the spatiotemporal risk field evolution model. The spatiotemporal risk field evolution model is used to iteratively solve the distribution and evolution trend of risk in the spatial and temporal dimensions, and output the four-dimensional spatiotemporal risk field under the current operation scenario. Based on the comparison results between the real-time field value of the four-dimensional spatiotemporal risk field and the multi-level security threshold, an access control strategy bound to the current risk level is obtained and executed. The access control strategy includes the linkage adjustment of the authentication factor combination method, reagent unlocking range and monitoring sampling frequency.

[0006] On one hand, a computer device is provided, the computer device including one or more processors and one or more memories, the one or more memories storing at least one computer program, the computer program being loaded and executed by the one or more processors to implement the method for batch access and access control of multiple reagents in the hazardous chemicals cabinet.

[0007] On the one hand, a computer-readable storage medium is provided, wherein at least one computer program is stored in the computer-readable storage medium, the computer program being loaded and executed by a processor to implement the method for batch access and access control of multiple reagents in the hazardous chemical cabinet.

[0008] On the one hand, a computer program product or computer program is provided, which includes program code stored in a computer-readable storage medium. The processor of a computer device reads the program code from the computer-readable storage medium and executes the program code, causing the computer device to execute the above-mentioned method for batch retrieval and access control of multiple reagents in a hazardous chemical cabinet. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1This is a schematic diagram of the implementation environment of a method for batch dispensing of multiple reagents and authorization linkage in a hazardous chemical cabinet, as provided in an embodiment of this application. Figure 2 This is a flowchart of a method for batch retrieval of multiple reagents and permission linkage in a hazardous chemical cabinet, provided in an embodiment of this application. Figure 3 This is a flowchart of another method for batch retrieval of multiple reagents and permission linkage in a hazardous chemical cabinet, provided in an embodiment of this application. Figure 4 This is a flowchart of another method for batch retrieval of multiple reagents and permission linkage in a hazardous chemical cabinet, provided in an embodiment of this application. Figure 5 This is a flowchart of another method for batch retrieval of multiple reagents and permission linkage in a hazardous chemical cabinet, provided in an embodiment of this application. Figure 6 This is a schematic diagram of a system for batch dispensing and access control of multiple reagents in a hazardous chemical cabinet, as provided in an embodiment of this application. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0012] In this application, the terms "first," "second," etc., are used to distinguish identical or similar items with essentially the same function. It should be understood that there is no logical or temporal dependency between "first," "second," and "nth," nor are there any restrictions on quantity or execution order.

[0013] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, data stored, data displayed, etc.) and signals involved in this application are all authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0014] Figure 1 This is a schematic diagram illustrating the implementation environment of a method for batch dispensing and access control of multiple reagents in a hazardous chemicals cabinet, as provided in an embodiment of this application. (See attached diagram.) Figure 1 The implementation environment may include node 110 and system 140.

[0015] Node 110 is connected to system 140 via a wireless or wired network. Optionally, node 110 is configured inside a hazardous chemical cabinet, and node 110 has an application installed and running that supports batch access to multiple reagents from the hazardous chemical cabinet and permission linkage.

[0016] System 140 is a standalone physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, Content Delivery Network (CDN), and big data and artificial intelligence platforms. System 140 can provide background services for applications running on node 110.

[0017] Traditional intelligent management cabinets for hazardous chemicals primarily rely on identification and monitoring of the physical state of individual reagent bottles for control. However, these technologies cannot detect the risk of chemical incompatibility coupling between different reagents in scenarios involving the bulk dispensing of multiple reagents. Furthermore, the fixed preset access control strategies are disconnected from real-time risk status and cannot be dynamically adjusted to effectively prevent potentially dangerous operations. In addition, the time dimension is not fully utilized for risk trend analysis and prediction.

[0018] To address this issue, this application proposes a method for batch access and access control of multiple reagents from hazardous chemical storage cabinets. See [link to relevant documentation]. Figure 2 The method includes the following steps: 201. In response to a bulk retrieval request containing at least two reagents to be retrieved, retrieve the chemical kinetic parameters corresponding to the reagents to be retrieved from a pre-stored chemical characteristic parameter library. The chemical kinetic parameters include the activation energy and heat of reaction. The activation energy is corrected by applying a temperature parameter from real-time environmental monitoring data, thereby obtaining a temperature-corrected activation energy parameter.

[0019] 202. Based on temperature-corrected activation energy parameters, reaction heat, and real-time environmental monitoring data, the coupling risk coefficients for at least two reagents to be used are calculated in real time for pairwise combinations. The calculation results for all reagent combinations are aggregated to generate a multi-reagent coupling risk matrix. This multi-reagent coupling risk matrix is ​​used to quantitatively characterize the chemical incompatibility risk of all reagent combinations within the batch use list under real-time conditions.

[0020] 203. Input the multi-reagent coupled risk matrix and the cabinet spatial location data obtained from the sensor network into the spatiotemporal risk field evolution model. Iteratively solve the distribution and evolution trend of risk in the spatial and temporal dimensions through the spatiotemporal risk field evolution model, and output the four-dimensional spatiotemporal risk field under the current operating scenario.

[0021] 204. Based on the comparison results between the real-time field value of the four-dimensional spatiotemporal risk field and the multi-level security thresholds, an access control strategy bound to the current risk level is obtained and executed. This access control strategy includes the linkage adjustment of the authentication factor combination method, reagent unlocking range, and monitoring sampling frequency.

[0022] For ease of understanding, the following explains some key terms in this embodiment: A chemical characteristic parameter library is configured to store the inherent chemical properties of various chemical reagents, such as reaction activation energy, heat of reaction, pre-exponential factor, gas constant, temperature-corrected baseline parameter set, nonlinear coupling coefficient, synergistic amplification factor, and cross-coupling compensation matrix. These parameters form the basis for assessing the potential chemical reaction risks between reagents.

[0023] Chemical kinetic parameters are a set of parameters used to describe the rate and mechanism of chemical reactions. They mainly include the activation energy and heat of reaction. The activation energy characterizes the energy barrier for a chemical reaction to occur, while the heat of reaction characterizes the absorption or release of energy during the chemical reaction.

[0024] The multi-reagent coupling risk matrix is ​​used to quantify the chemical incompatibility risk of all reagent combinations within a batch purchase list under real-time conditions. Each element of the matrix corresponds to a coupling risk coefficient for a pair of reagent combinations, comprehensively reflecting the reaction tendency, reaction intensity, and the influence of environmental factors between reagents.

[0025] The spatiotemporal risk field evolution model is designed to simulate and predict the distribution and evolution trends of risks within hazardous chemical containers in both spatial and temporal dimensions. This model iteratively solves for the risk state by considering factors such as risk source terms, spatial propagation, and temporal decay.

[0026] The four-dimensional spatiotemporal risk field is the output of the spatiotemporal risk field evolution model. It is used to characterize the real-time distribution and evolution of risks within hazardous chemical containers in three-dimensional space and time under the current operational scenario. Its field value reflects the risk intensity at a specific spatiotemporal point.

[0027] Access control policies are a set of rules used to dynamically adjust user access permissions based on real-time risk levels. These policies can include adjustments to authentication factor combinations, reagent unlocking ranges, and monitoring sampling frequencies to meet security needs at different risk levels.

[0028] The method provided in this embodiment allows users to manually input the name or number of the reagents to be taken via an interface when responding to a batch request containing at least two reagents. The system performs a precise match in a chemical characteristic parameter library based on the received input information, thereby directly obtaining preset activation energy and reaction heat data. Further, the system obtains the current environmental temperature parameter from real-time environmental monitoring data. This real-time temperature parameter is then linearly interpolated with the activation energy obtained from the chemical characteristic parameter library. Alternatively, a preset temperature correction lookup table can be consulted to directly obtain the activation energy correction value corresponding to the current temperature. Thus, the temperature-corrected activation energy parameter is obtained.

[0029] When calculating the coupling risk coefficient of at least two reagents to be used in pairwise combinations in real time based on temperature-corrected activation energy parameters, heat of reaction, and real-time environmental monitoring data, a preliminary coupling risk value can be calculated for each pair of reagents in the batch usage list, according to its temperature-corrected activation energy parameters and heat of reaction, combined with real-time environmental monitoring data (such as humidity, air pressure, etc.), through a preset simple weighted summation model or linear regression model. These preliminary coupling risk values ​​are organized into a matrix structure, thus forming a multi-reagent coupling risk matrix. Each element of this matrix directly reflects the incompatibility risk of the corresponding reagent combination under the current environment.

[0030] When inputting the multi-reagent coupled risk matrix along with cabinet spatial location data acquired from the sensor network into the spatiotemporal risk field evolution model, the generated multi-reagent coupled risk matrix can be used as the initial risk input, combined with the static spatial coordinate information of each storage location acquired from the cabinet sensor network. This data is then fed into a simplified spatiotemporal risk evolution model. This model can employ rule-based propagation logic, for example, setting the risk value to diffuse between adjacent storage locations at a fixed ratio and decay linearly at a fixed rate over time. Through a finite number of simulations based on these rules, the model outputs a four-dimensional data structure containing predicted risk values ​​at different spatial locations and several future time steps—that is, a four-dimensional spatiotemporal risk field.

[0031] When comparing the real-time field value of the four-dimensional spatiotemporal risk field with multi-level security thresholds to obtain and execute the access control policy bound to the current risk level, the real-time risk field value at the current moment can be extracted from the four-dimensional spatiotemporal risk field. These real-time field values ​​are directly compared with preset fixed multi-level security thresholds (e.g., low-risk threshold, medium-risk threshold, high-risk threshold). Based on the comparison results, the system determines the current risk level and directly retrieves the access control policy bound to that risk level from a predefined access control policy lookup table. This policy is then issued to the execution module to adjust the authentication factor combination method (e.g., from single-factor authentication to two-factor authentication), reagent unlocking range (e.g., restricting the types or quantities of reagents that can be unlocked), and monitoring sampling frequency (e.g., from sampling once per minute to sampling once every 10 seconds).

[0032] The method provided in this application constructs a four-dimensional spatiotemporal risk field by real-time sensing of chemical incompatibility coupling risks in multi-reagent batch dispensing scenarios and combining this with spatial location data within the cabinet. This enables a quantitative assessment of the distribution and evolution trends of risks in both spatial and temporal dimensions. Consequently, access control strategies can be dynamically adjusted based on real-time risk levels, including authentication factor combinations, reagent unlocking ranges, and monitoring sampling frequencies. This method overcomes the shortcomings of related technologies in multi-reagent risk sensing and rigid access control, thereby improving the safety management level of hazardous chemical cabinets.

[0033] In some of the solutions described above in this application, chemical kinetic parameters are retrieved from a chemical characteristic parameter library to calculate the coupling risk coefficient. However, in this process, when there is a lack of directly related reaction activation energy data between any two reagents to be retrieved, complete chemical kinetic parameters cannot be obtained, resulting in inaccurate risk calculation or inability to perform the calculation.

[0034] In response to this, this application further proposes a method that, in response to a batch retrieval request containing at least two reagents to be retrieved, retrieves the chemical kinetic parameters corresponding to the reagents to be retrieved from a pre-stored chemical feature parameter library. This includes: in response to a batch retrieval request containing at least two reagents to be retrieved, identifying the reagent identifiers and reagent categories of the at least two reagents to be retrieved based on the batch retrieval request; querying the basic activation energy and basic heat of reaction corresponding to each reagent to be retrieved from the chemical feature parameter library based on the reagent identifiers and reagent categories; when no directly correlated activation energy data is found between any two reagents to be retrieved, using the reagent categories of the two reagents to be retrieved, calling the pre-stored group contribution estimation rules in the chemical feature parameter library, and combining the molecular structure characteristics of the two reagents to be retrieved to generate an estimated activation energy; and integrating the basic activation energy, the basic heat of reaction, and the estimated activation energy to generate a chemical kinetic parameter set corresponding to the batch retrieval request.

[0035] Specifically, when identifying the reagent identifiers and corresponding reagent categories of at least two reagents to be dispensed, the purpose is to clarify the specific reagents and their chemical classifications involved in the bulk dispensing request, providing basic information for subsequent parameter queries and estimations. One implementation method is to automatically identify the reagent identifier (e.g., CAS number, product batch number) by scanning the barcode or QR code on the reagent bottle and match it with its preset reagent category (e.g., acid, alkali, oxidizing agent, reducing agent, organic solvent, etc.) from the internal database. Another implementation method is for the user to manually enter the reagent name or number on the bulk dispensing request interface, and the system retrieves the corresponding reagent identifier and reagent category from the reagent management system through fuzzy matching or precise querying.

[0036] When querying the basic activation energy and basic heat of reaction corresponding to each reagent to be used from this chemical characteristic parameter database, its role is to obtain known and reliable chemical reaction kinetic data as a direct basis for risk assessment. This chemical characteristic parameter database can be a structured database storing a large amount of known reagent pairs' activation energies and heats of reaction data. The system performs precise matching queries directly based on the identified reagent identifiers. Alternatively, this chemical characteristic parameter database can also be a knowledge graph, where there are relationships between entities such as reagents, reaction types, activation energies, and heats of reaction. The system uses a graph query algorithm to retrieve basic reaction parameters related to the reagent to be used.

[0037] When no direct correlation data for the activation energy of any two reagents is found, the system uses pre-stored group contribution estimation rules from the chemical feature parameter library, based on the reagent categories of these two reagents, and combines this with the molecular structural features of the reagents to generate an estimated activation energy. This step aims to address the problem of missing data, estimating the activation energy of unknown reactions through chemical principles and ensuring the completeness of the risk assessment. The group contribution method is a method that estimates the overall properties of a molecule by breaking it down into functional groups and considering the contributions of these functional groups to specific properties. When direct data is missing, the system selects the appropriate group contribution model based on the reagent category (e.g., whether it is organic or inorganic). Then, by analyzing the molecular structure of the reagents (e.g., through chemical structural formula analysis), key functional group information is extracted, and pre-stored functional group contribution values ​​are summed or weighted to estimate the activation energy. Besides the group contribution method, the estimation rules can also be based on machine learning models trained on a large amount of known activation energy data, learning the mapping relationship between reagent category, molecular structural features, and activation energy. When data is missing, the type and molecular structure characteristics of the reagent to be used are used as input, and the activation energy is estimated by model prediction.

[0038] The system integrates the basic activation energy, basic heat of reaction, and estimated activation energy to generate a set of chemical kinetic parameters corresponding to the batch retrieval request. This integration process aims to unify and standardize all acquired chemical kinetic parameters, providing a consistent data interface for subsequent coupling risk calculations. The system organizes and stores all queried basic activation energies and heats of reaction, as well as all activation energies obtained through estimation methods, according to a predetermined data structure (e.g., a two-dimensional matrix or a list), ensuring that each reagent pair has corresponding activation energy and heat of reaction data. The integration process may also include data cleaning and validation. For example, for certain reagent pairs, if both basic and estimated data exist, the system can select based on a preset priority (e.g., basic data priority) or perform a weighted average to generate the parameter set.

[0039] Through the above technical solution, this application effectively addresses the problem of inaccurate or incomplete risk assessment caused by data gaps in the chemical characteristic parameter library during batch dispensing of multiple reagents. By identifying reagent identifiers and categories, the system can accurately query known basic reaction activation energies and heats of reaction. More importantly, when faced with unknown or data-missing reagent combinations, the system can intelligently invoke pre-stored group contribution estimation rules, combined with the molecular structural characteristics of the reagents, to generate reliable estimated activation energies. This mechanism ensures that even with incomplete data, a complete set of chemical kinetic parameters can be obtained, providing a comprehensive and accurate data foundation for subsequent real-time calculation of coupling risk coefficients. This avoids interruptions or inaccuracies in risk assessment due to data gaps, improving the completeness and reliability of risk perception in batch dispensing scenarios involving multiple reagents in hazardous chemical cabinets.

[0040] In some embodiments described above in this application, the group contribution method is proposed to estimate the activation energy when directly related activation energy data is missing. However, in its implementation, a specific method is needed to accurately generate the estimated activation energy based on molecular structure characteristics to ensure the reliability of the risk calculation and avoid inaccurate quantification of chemical incompatibility risk due to missing data.

[0041] To address this, this application further proposes a method for estimating activation energy based on the reagent categories of any two reagents to be used, by invoking pre-stored group contribution estimation rules in a chemical characteristic parameter library, and combining the molecular structural characteristics of any two reagents to be used. This method includes: obtaining the molecular structural formulas of any two reagents to be used from a batch request; performing functional group decomposition on the molecular structural formulas to extract multiple functional group fragments contained in each of the two reagents to be used; querying multiple group contribution values ​​corresponding to the multiple functional group fragments from the chemical characteristic parameter library based on the reagent categories of the two reagents to be used, whereby these group contribution values ​​are used to quantify the contribution of a single functional group fragment to the reaction activation energy; and performing a weighted summation of the multiple group contribution values ​​to obtain the estimated activation energy between any two reagents to be used.

[0042] Specifically, when obtaining the molecular structural formulas of any two reagents to be used, this step aims to ensure the acquisition of accurate chemical structural information of the reagents, which is the basis for subsequent functional group decomposition and group contribution estimation. The molecular structural formula, as a unique identifier of a chemical substance, directly determines its chemical properties. For example, after receiving a batch purchase request, the system can retrieve the corresponding molecular structural formula from an internal database or external chemical information platform based on the reagent identifier (such as CAS number, chemical name, etc.) included in the request, such as the SMILES string, InChI code, or two-dimensional / three-dimensional structural diagram data. Furthermore, when submitting a batch purchase request, users can also directly upload or input the molecular structural formula data of the reagents, which the system receives and parses.

[0043] When decomposing a molecular structure into functional group fragments to extract multiple functional group segments from any two reagents, a functional group is an atom or group of atoms in the molecule that determines its chemical properties. Decomposing a molecular structure into functional group fragments simplifies complex molecular structures into quantifiable basic units, enabling activation energy estimation using the group contribution method. For example, a rule-based algorithm can be used, with a pre-defined functional group recognition rule base. Pattern matching (such as SMARTS patterns) can scan and identify the input molecular structure, decomposing the molecule into known functional group fragments and the remaining skeleton. Alternatively, a machine learning model can be used. By training on a large number of known molecular structures and their corresponding functional group relationships, the model can automatically identify and extract the main functional group fragments in the molecule, and even identify undefined fragments with specific chemical activities.

[0044] When querying the chemical feature parameter library for multiple group contribution values ​​corresponding to multiple functional group fragments based on any two reagent categories to be used, these group contribution values ​​are used to quantify the contribution of a single functional group fragment to the activation energy of the reaction. Querying by reagent category ensures that the obtained group contribution values ​​are specific to the reaction type (e.g., redox reaction, acid-base reaction, etc.) and reagent environment, thereby improving the accuracy of the estimation. For example, the chemical feature parameter library can pre-store group contribution values ​​for various common functional groups under different reagent categories. The system performs an index query in the parameter library based on the category of the reagent to be used (e.g., "oxidizing agent," "reducing agent," "acid," "base," etc.) and the decomposed functional group fragments to obtain the corresponding values. Furthermore, the chemical feature parameter library can also be a multidimensional database, where group contribution values ​​are not only associated with functional groups and reagent categories, but may also be associated with other parameters such as reaction type and temperature range. The system uses multi-condition matching to accurately retrieve the group contribution value that best matches the current scenario.

[0045] When calculating the estimated activation energy between any two reagents by weighted summation of the contributions of multiple functional groups, the weighted summation combines the contributions of different functional groups according to their importance or number in the molecule to obtain an estimated activation energy for the entire molecule or intermolecular reaction. This is a common method for estimating chemical properties and can reflect the overall reactivity of the molecule. For example, a simple weighted summation can be used, which directly adds the contributions of all identified functional group fragments, or weights can be applied based on the frequency of occurrence of the functional group in the molecule. Furthermore, more complex weighted models can be used, such as considering steric hindrance, electronic effects, or conjugation effects between functional groups, assigning different weighting factors to different functional groups, and then performing a weighted summation. These weighting factors can be predetermined based on experimental data or quantum chemical calculations.

[0046] Through the above technical solution, this application provides a scientific and reliable method for estimating activation energy in the absence of direct experimental data. Specifically, by obtaining the molecular structure formula of the reagent to be retrieved from the batch retrieval request, the accuracy of the basic data for estimation is ensured. By decomposing the functional groups of the molecular structure formula, the complex molecular structure is simplified into tractable units, improving the universality of the estimation. Based on the reagent category of the reagent to be retrieved, the contribution value of the functional group corresponding to the fragment is queried from the chemical characteristic parameter library, making the estimation results more targeted and accurate, avoiding blind estimation. By weighted summation of the contribution values ​​of multiple functional groups, the contribution of each functional group is integrated, providing a reliable estimated activation energy. This effectively solves the problem of inaccurate quantification of chemical incompatibility risk due to data deficiency, thereby ensuring the accurate construction of the multi-reagent coupling risk matrix, and thus improving the risk perception accuracy and the rationality of access control of the multi-reagent batch retrieval and access control method for hazardous chemical cabinets.

[0047] In some embodiments described above in this application, a method is proposed to apply a real-time temperature parameter to the reaction activation energy to correct it in order to calculate the risk coefficient more accurately. However, in its implementation, due to the nonlinear relationship between temperature and activation energy, a simple correction method may lead to inaccurate correction and fail to truly reflect the impact of real-time temperature changes on the reaction rate, thereby affecting the accuracy of the subsequent risk matrix.

[0048] To address this, this application further proposes correcting the activation energy of the reaction by applying the temperature parameter from real-time environmental monitoring data, resulting in a temperature-corrected activation energy parameter, see [link to relevant documentation]. Figure 3 The correction method includes the following steps: 301. Extract the real-time temperature data for the current operating scenario from the real-time environmental monitoring data.

[0049] 302. Obtain the basic reaction activation energy corresponding to the reagent to be used from the chemical characteristic parameter library.

[0050] 303. Input the real-time temperature data and the activation energy of the basic reaction into the temperature correction model. Use the temperature correction model to fit and calculate the nonlinear relationship between temperature and activation energy, and output the temperature correction coefficient.

[0051] 304. Multiply the temperature correction factor by the activation energy of the basic reaction to obtain the temperature-corrected activation energy parameter.

[0052] Specifically, real-time temperature data under the current operating scenario is extracted from this real-time environmental monitoring data to provide real-time environmental input for subsequent activation energy correction, ensuring the timeliness and accuracy of the correction results. This real-time temperature data can be collected in real time by temperature sensors (e.g., thermistors, thermocouples, infrared thermometers, etc.) deployed inside the hazardous materials container or operating area, and the collected analog signals are converted into digital signals and transmitted to the processing unit through the data acquisition module. Alternatively, the environmental monitoring module integrated into the intelligent management system can periodically or on-demand acquire temperature data from the sensor network, and perform data cleaning and formatting to ensure data availability and consistency.

[0053] Obtaining the baseline activation energy of the reagent to be used from the chemical characteristic parameter library is to provide a pure chemical reaction characteristic parameter, unaffected by real-time temperature, as a starting point for correction. This baseline activation energy can be obtained by establishing a reagent-activation energy mapping table in the chemical characteristic parameter library. When the identifier of the reagent to be used is identified, the system automatically queries and extracts the activation energy value of that reagent under standard conditions. Alternatively, the activation energy data of different reagents at specific reference temperatures can be entered into the chemical characteristic parameter library through prior experimental determination or literature review of common hazardous chemical reagents, and retrieved through the database query interface when needed.

[0054] Inputting the real-time temperature data and the basic activation energy into the temperature correction model allows it to comprehensively consider current environmental conditions and the inherent properties of the reagents, enabling accurate activation energy correction calculations. This can be achieved through a programming interface (API) or data bus, where the real-time temperature data and the basic activation energy obtained from a chemical characteristic parameter library are encapsulated into a model-acceptable input format (e.g., JSON object, structure, or array) and passed to the temperature correction model's calculation module. Alternatively, a function or method can be defined at the software level that accepts the real-time temperature value and the basic activation energy value as parameters and internally calls the temperature correction model's algorithm for processing.

[0055] This temperature correction model fits and calculates the nonlinear relationship between temperature and activation energy, outputting a temperature correction coefficient. The aim is to accurately quantify the impact of temperature changes on activation energy by establishing or invoking a mathematical model that captures this nonlinear relationship, and outputting the result as a correction coefficient. For example, machine learning models (such as neural networks or support vector regression) can be trained on large amounts of experimental data to learn the nonlinear mapping between temperature and activation energy, thereby predicting the correction coefficient at a given real-time temperature. Alternatively, empirical or semi-empirical formulas based on physicochemical principles (such as variants of the Arrhenius equation) can be used to determine the parameters by fitting experimental data, and then the corresponding correction coefficient can be calculated based on the real-time temperature.

[0056] Multiplying the temperature correction factor by the base reaction activation energy yields the temperature-corrected activation energy parameter, which is the step in completing the activation energy temperature correction. This can be achieved through a simple numerical multiplication operation: temperature-corrected activation energy parameter = base reaction activation energy × temperature correction factor. Alternatively, in the software module, the base reaction activation energy can be used as a reference, and the equivalent activation energy parameter at the current real-time temperature can be obtained by applying the correction factor for proportional adjustment.

[0057] Through the above technical solution, this application effectively solves the problem that simple correction methods may ignore the nonlinear relationship between temperature and activation energy. Specifically, by extracting real-time temperature data under the current operating scenario from real-time environmental monitoring data, the real-time nature and situational adaptability of the correction process are ensured. Obtaining the basic reaction activation energy corresponding to the reagent to be used from the chemical characteristic parameter library provides a benchmark for correction. The real-time temperature data and the basic reaction activation energy are input into the temperature correction model, and this model is used to fit and calculate the nonlinear correlation between temperature and activation energy, outputting an accurate temperature correction coefficient. This overcomes the limitations of traditional linear correction, enabling the corrected activation energy parameter to truly and accurately reflect the impact of real-time temperature changes on the reaction rate. Multiplying this temperature correction coefficient by the basic reaction activation energy yields the temperature-corrected activation energy parameter. This parameter more accurately characterizes the chemical reactivity of the reagent in the current environment, providing a reliable input for the accurate calculation of the subsequent multi-reagent coupling risk matrix, and improving the risk perception accuracy and overall safety of the multi-reagent batch retrieval and access control method for hazardous chemical cabinets.

[0058] In some embodiments of this application, a temperature correction model is proposed to fit and calculate the nonlinear correlation between temperature and activation energy to output a temperature correction coefficient, which is used to correct the activation energy parameter. However, in its implementation, the fitting of the nonlinear correlation may lack a specific mechanism, which may cause the temperature correction coefficient to fail to accurately reflect the trend of activation energy change when the temperature deviates from the reference point, thus affecting the authenticity of the activation energy parameter.

[0059] In response, this application further proposes a step-by-step approach to fit and calculate the nonlinear relationship between temperature and activation energy using a temperature correction model, and to output the temperature correction coefficient: The temperature correction model retrieves the temperature correction reference parameter set corresponding to the reagent to be used from the chemical characteristic parameter library. The temperature correction reference parameter set includes reference temperature data and activation energy temperature sensitivity coefficient. The activation energy temperature sensitivity coefficient is used to quantify the degree of influence of temperature change on the activation energy of the reaction.

[0060] The temperature difference between the real-time temperature data and the reference temperature data is calculated using this temperature correction model.

[0061] The temperature difference and the activation energy temperature sensitivity coefficient are input into the exponential fitting function of the temperature correction model. The exponential fitting function is used to perform a nonlinear mapping on the activation energy change trend when the temperature deviates from the reference temperature, and the temperature correction coefficient is obtained.

[0062] The temperature correction model is a computational module used to adjust chemical kinetic parameters (especially activation energy) to adapt to real-time temperature changes. This model can be a software algorithm integrated into a hazardous materials storage management system, responsible for data querying, mathematical calculations, and result output. Alternatively, it can be a standalone microservice that interacts with the main system via an API interface to perform specialized temperature correction. The chemical characteristic parameter library is a structured data storage system used to store the inherent properties and reaction kinetic parameters of various chemical reagents. This library can use a relational database (such as MySQL or PostgreSQL) to store reagent identification, molecular structure, reaction type, and corresponding basic reaction activation energy, heat of reaction, pre-exponential factor, etc., and includes a baseline parameter set for temperature correction. Alternatively, it can be a distributed file system that stores chemical data in JSON or XML format and manages and queries it through a specific data access layer. The temperature correction baseline parameter set is a dataset specifically for temperature correction within the chemical characteristic parameter library, providing the foundation for temperature-dependent correction of activation energy. This parameter set can store a standard reference temperature (e.g., 25°C or 298.15 K) and an activation energy temperature sensitivity coefficient for each reagent or reagent combination to be used. Alternatively, it can contain a series of activation energy data measured at different temperature points for model interpolation or more complex fitting. The reference temperature data is a preset or experimentally determined baseline temperature value used to measure the degree of deviation from the real-time temperature. This reference temperature data can be an industry standard temperature, such as 25°C commonly used in chemical reaction kinetics studies. Alternatively, it can be the experimental temperature corresponding to the baseline activation energy of the reagent being measured. The activation energy temperature sensitivity coefficient is a quantitative indicator used to describe the degree to which the activation energy changes with temperature. This coefficient can be a constant obtained experimentally (e.g., Arrhenius plot analysis) or theoretically calculated (e.g., quantum chemical calculations), whose value reflects the intrinsic response characteristics of the activation energy of a specific reaction to temperature fluctuations. Alternatively, it can be a temperature-dependent function to more finely capture the complexity of activation energy changes with temperature. The real-time temperature data refers to the actual temperature measurement value under the current operating scenario obtained from the sensor network. This data can be collected in real time and transmitted to the system by temperature sensors (such as thermocouples or platinum resistance thermometers) installed inside the hazardous materials container. Alternatively, it can be a representative temperature value obtained by averaging or weighting data from multiple temperature sensors inside the container. The temperature difference is the numerical difference between the real-time temperature data and the reference temperature data. This difference is obtained through a simple subtraction operation, i.e., "real-time temperature data - reference temperature data," with the sign indicating whether the real-time temperature is higher or lower than the reference temperature. The exponential fitting function is a mathematical model used to describe the nonlinear relationship between activation energy and temperature.This function can take the classic exponential form, such as f(ΔT) = A × exp(B × ΔT), where ΔT is the temperature difference, and A and B are parameters related to the activation energy temperature sensitivity coefficient. Alternatively, it can be a variant based on the Arrhenius or Eyring equations, specifically designed to fit the temperature dependence of the activation energy. This nonlinear mapping refers to the process of converting the temperature difference and activation energy temperature sensitivity coefficient into a temperature correction factor through an exponential fitting function. This process calculates a nonlinear correction factor that accurately reflects the effect of temperature deviation on the activation energy by substituting the input parameters into the exponential fitting function. Alternatively, it can be implemented using a pre-built lookup table that stores correction factors for different combinations of temperature differences and sensitivity coefficients to improve computational efficiency. This temperature correction factor is a multiplicative factor used to adjust the activation energy of the basic reaction. This coefficient is the output of the exponential fitting function, and its value can be greater than 1 (indicating an increase in activation energy) or less than 1 (indicating a decrease in activation energy), thus correcting the basic activation energy to the actual activation energy at the current real-time temperature.

[0063] Through the above technical solution, this application solves the problem of inaccurate nonlinear fitting by refining the operation process of the temperature correction model, ensuring that the temperature correction coefficient can accurately quantify the impact of temperature changes on the activation energy. Specifically, the temperature correction model retrieves a temperature correction benchmark parameter set from a chemical characteristic parameter library. This parameter set includes reference temperature data and the activation energy temperature sensitivity coefficient, which establishes a benchmark framework for fitting, enabling the model to quantify the degree of influence of temperature changes based on pre-stored parameters and avoid relying on simple assumptions. The temperature correction model calculates the temperature difference between real-time temperature data and reference temperature data, which directly captures the deviation between the current environment and standard conditions, providing accurate input for nonlinear mapping. Then, the temperature difference and the activation energy temperature sensitivity coefficient are input into an exponential fitting function, which performs a nonlinear mapping on the trend of activation energy change when the temperature deviates, generating the temperature correction coefficient. Using an exponential fitting function can effectively handle the nonlinear relationship between activation energy and temperature, avoid the limitations of linear models, and ensure that the correction coefficient truly reflects the dynamic changes of activation energy under temperature fluctuations, thereby improving the accuracy of subsequent risk calculations.

[0064] In some of the solutions mentioned above in this application, a multi-reagent coupling risk matrix is ​​generated by real-time calculation of the coupling risk coefficient based on temperature-corrected activation energy parameters, reaction heat, and real-time environmental monitoring data. This matrix is ​​used to quantify the chemical incompatibility risk of all reagent combinations in the batch use list under real-time conditions. However, in this process, there is a lack of detailed calculation mechanisms for thermodynamic tendency, comprehensive intensity of energy release, synergistic contribution of actual reactant mass and contact probability, and cross-coupling effects between multiple reagents, resulting in insufficient accuracy in risk quantification.

[0065] To address this, this application further proposes a method for real-time calculation of the coupling risk coefficients for at least two reagents to be used, based on temperature-corrected activation energy parameters, reaction heat, and real-time environmental monitoring data. The calculation results for all reagent combinations are then aggregated to generate a multi-reagent coupling risk matrix. (See [link to relevant documentation]). Figure 4 ,include: 401. Perform an exponential domain transformation on the temperature-corrected activation energy parameter to obtain the reaction potential factor corresponding to each reagent pair. The reaction potential factor is used to characterize the thermodynamic tendency of the chemical reaction to occur under the current temperature conditions.

[0066] 402. The reaction potential factor and the corresponding reaction heat are nonlinearly coupled to obtain the reaction intensity base value for each reagent combination. The reaction intensity base value is used to quantify the comprehensive intensity of the chemical reaction in both thermodynamic driving and energy release dimensions.

[0067] 403. Extract real-time concentration data and remaining amount data corresponding to at least two reagents to be taken from real-time environmental monitoring data, and map the real-time concentration data and remaining amount data together as a dynamic weight of reaction probability. The dynamic weight of reaction probability is used to characterize the synergistic contribution of actual reactant mass and reaction contact probability to the degree of risk.

[0068] 404. Based on the reaction intensity base value, the dynamic weight of the reaction probability, and the cross-coupling compensation mechanism among multiple reagents, the coupling risk coefficient is comprehensively calculated and matrixed to generate a multi-reagent coupling risk matrix. The cross-coupling compensation mechanism is used to quantify the catalytic or inhibitory effect of the third type of reagent among at least two reagents to be used on the current reagent combination.

[0069] Specifically, the reaction potential factor is obtained by performing an exponential domain transformation on the temperature-corrected activation energy parameter. The exponential domain transformation is a technique that maps activation energy parameters from linear space to exponential space. Its core lies in using mathematical functions to transform minute changes in activation energy into differences in their impact on reaction rate or tendency. For example, an exponential function based on the Arrhenius or Eyring equations can be used, where the temperature-corrected activation energy parameter, pre-exponential factor, gas constant, and real-time temperature data are input into the exponential transformation function, and the reaction potential factor is obtained through negative exponential calculation. Alternatively, other nonlinear mapping functions, such as the Sigmoid or Tanh functions, can be used to map the activation energy parameter to probability values ​​between 0 and 1 to characterize the thermodynamic tendency of the reaction. The reaction potential factor is a quantitative index obtained after the exponential domain transformation; it directly reflects the ease with which a chemical reaction overcomes the activation energy barrier and the thermodynamic driving force under the current temperature conditions. A larger value indicates a stronger thermodynamic tendency for the reaction to occur.

[0070] The reaction intensity baseline is obtained by nonlinearly coupling the reaction potential factor with the corresponding reaction heat. Nonlinear coupling refers to the fusion calculation of these two key parameters—reaction potential factor and reaction heat—through a nonlinear mathematical model to more accurately reflect the overall intensity of a chemical reaction. This method can capture potential synergistic amplification or inhibition effects between the two, rather than a simple linear superposition. For example, a power-law coupling function can be used, taking the product of the reaction potential factor and reaction heat as the base and introducing a nonlinear coupling coefficient as the exponent or multiplier, thereby fitting a nonlinear amplification relationship between reaction potential and reaction heat to obtain an intensity amplification factor, which is then multiplied by the intermediate value of the linear coupling. Alternatively, polynomial functions, logarithmic functions, or neural network-based models can be used to establish a complex nonlinear relationship between the reaction potential factor and reaction heat by learning from historical reaction data, thus outputting the reaction intensity baseline. The reaction intensity baseline is a comprehensive intensity index that quantifies the chemical reaction in two dimensions: thermodynamic driving force (characterized by the reaction potential factor) and energy release (characterized by the reaction heat). It comprehensively considers the probability of the reaction occurring and the potential destructive extent once it occurs.

[0071] Real-time concentration and remaining quantity data corresponding to at least two reagents to be used are extracted from real-time environmental monitoring data, and these data are jointly mapped to dynamic weights for the reaction probability. Mapping these data to dynamic weights aims to comprehensively consider the accessibility and reactivity of actual reactants in the current environment, thereby more accurately assessing the actual probability of the reaction occurring. This mapping process typically involves normalization and weighted fusion of the raw data. For example, real-time concentration data can be normalized to obtain a concentration contribution coefficient, and remaining quantity data can be normalized to obtain a mass contribution coefficient. These two coefficients are then multiplied by a co-amplification factor retrieved from a chemical feature parameter library to obtain the dynamic weights for the reaction probability. Alternatively, fuzzy logic or machine learning-based methods can be used, taking real-time concentration and remaining quantity as input, and dynamically outputting a probability value between 0 and 1 as the dynamic weights for the reaction probability through a pre-set rule set or a trained model. The dynamic weight of reaction probability is a dynamically changing quantity that comprehensively reflects the synergistic contribution of actual reactant mass (such as residual amount) and reaction contact probability (such as concentration) to the degree of risk of chemical reaction occurrence, ensuring that risk assessment considers not only intrinsic chemical properties, but also the state of matter under actual operating conditions.

[0072] Based on the base value of reaction intensity, dynamic weights of reaction probability, and a cross-coupling compensation mechanism among multiple reagents, a comprehensive solution and matrix aggregation of coupling risk coefficients are performed to generate a multi-reagent coupling risk matrix. The cross-coupling compensation mechanism is used to quantify and adjust the catalytic or inhibitory effects of "third-class" reagents (besides the two directly interacting reagents) on the reaction risk of the current reagent combination in a multi-reagent system. This mechanism can capture potential chain reactions or mutual influences in complex multi-component systems. For example, a cross-coupling compensation matrix can be pre-constructed, where each compensation element quantifies the catalytic or inhibitory effect of the third-class reagent on a specific reagent combination, and then the base coupling risk value is multiplied element-wise by the corresponding compensation element. Alternatively, a compensation factor for a specific reagent combination can be dynamically calculated based on an expert experience rule base or machine learning model, according to the category and quantity of all reagents in the batch retrieval list. This compensation factor can be a value greater than 1 (representing a catalytic amplification effect) or a value less than 1 (representing an inhibitory attenuation effect). The comprehensive calculation and matrix aggregation of coupling risk coefficients refers to the systematic integration of the calculated reaction intensity baseline, reaction probability dynamic weights, and the quantitative results of cross-coupling compensation mechanisms to form a multi-reagent coupling risk matrix that can comprehensively characterize the risks of all reagent combinations. For example, an initial coupling risk matrix can be constructed based on a batch usage list. The reaction intensity baseline for each reagent combination is multiplied by the reaction probability dynamic weight to obtain a basic coupling risk value, which is then filled into the corresponding position in the initial coupling risk matrix. The filled initial coupling risk matrix is ​​then multiplied element-wise with the cross-coupling compensation matrix to amplify the catalytic effect or attenuate the inhibitory effect on the basic coupling risk value, resulting in a multi-reagent coupling risk matrix. Alternatively, a multi-factor weighted summation model can be used, assigning different weights to the reaction intensity baseline, reaction probability dynamic weights, and compensation factors, then performing a weighted summation to obtain coupling risk coefficients, and finally filling these coefficients into the matrix according to the reagent combination relationships. The multi-reagent coupling risk matrix is ​​a two-dimensional or multi-dimensional data structure in which each element represents the chemical incompatibility risk of a specific reagent combination within the batch retrieval list in a real-time environment. In matrix form, all potential coupling risks can be quantified and displayed intuitively and comprehensively.

[0073] Through the above technical solutions, this application can more accurately quantify the chemical incompatibility risk between reagents in scenarios involving the batch use of multiple reagents. Specifically, by performing an exponential domain transformation on the temperature-corrected activation energy parameter, the thermodynamic tendency of the chemical reaction at the current temperature can be accurately captured, making the risk assessment closer to the actual reaction conditions. Nonlinearly coupling the reaction potential factor and reaction heat comprehensively considers both the probability of the reaction occurring and the destructiveness of energy release, avoiding errors that may arise from linear models, thus obtaining a more representative baseline value for reaction intensity. Simultaneously, introducing real-time concentration data and residual data and mapping them to dynamic weights of reaction probability allows the risk assessment to dynamically reflect the synergistic contribution of actual reactant mass and contact probability to the degree of risk, avoiding the limitations of assessments based solely on theoretical parameters. Furthermore, by introducing a cross-coupling compensation mechanism between multiple reagents, the catalytic or inhibitory effect of a third type of reagent on the current reagent combination can be quantified, thus comprehensively considering all potential interactions in complex multi-component systems and avoiding the omission of key risk points. By comprehensively calculating and matrixing the coupled risk coefficients, the generated multi-reagent coupled risk matrix can comprehensively and intuitively characterize the chemical incompatibility risk of all reagent combinations in the batch retrieval list under real-time conditions. This provides more accurate risk source terms for subsequent spatiotemporal risk field evolution models, thereby improving the refinement and accuracy of risk perception of hazardous chemical cabinets and effectively reducing potential risks in batch retrieval operations of multiple reagents.

[0074] In some embodiments described above in this application, an exponential domain transformation of the temperature-corrected activation energy parameter is proposed to obtain a reaction potential factor to characterize the thermodynamic tendency of a chemical reaction. However, if the influence of molecular collision frequency and real-time temperature data is not fully considered during its implementation, the calculation of the reaction potential factor may not be accurate enough.

[0075] In this regard, this application further proposes to perform an exponential domain transformation on the temperature-corrected activation energy parameter to obtain the reaction potential factor corresponding to each reagent pair. This process includes: The pre-exponential factor and gas constant corresponding to each reagent pair are retrieved from the chemical characteristic parameter library. The pre-exponential factor is used to characterize the effect of molecular collision frequency on reaction rate.

[0076] The temperature-corrected activation energy parameter, the pre-exponential factor, the gas constant, and the real-time temperature data extracted from the real-time environmental monitoring data are all input into the exponential transformation function.

[0077] The ratio of the temperature-corrected activation energy parameter to the gas constant and the real-time temperature data is calculated using the exponential transformation function to obtain the intermediate value of the exponential calculation corresponding to each reagent pair.

[0078] Multiply the intermediate value of the exponent by the pre-exponential factor to obtain the reaction potential factor corresponding to each reagent pair.

[0079] Specifically, the pre-exponential factor and gas constant corresponding to each reagent pair are retrieved from the chemical characteristic parameter library. The pre-exponential factor characterizes the effect of molecular collision frequency on the reaction rate. Its concept originates from the Arrhenius equation and represents the frequency of effective collisions between reactant molecules per unit time, reflecting the spatial orientation and collision frequency characteristics of reactant molecules. In practical applications, the pre-exponential factor can be obtained in various ways. For example, it can be obtained by accurately measuring the reaction rate constant at different temperatures for a specific reagent pair through experimental kinetic measurements and fitting it to the linearized form of the Arrhenius equation. Alternatively, for reagents with similar structures or well-defined reaction mechanisms, quantum chemical calculation methods can be used to simulate the reaction pathway and transition state structure, thereby theoretically calculating the pre-exponential factor. These data are usually pre-stored in the chemical characteristic parameter library for system query use. The gas constant is a general physical constant, usually denoted by the symbol R, and its function is to correlate energy units with temperature units in chemical kinetic calculations. The value of the gas constant is fixed, for example, approximately 8.314 J / (mol·K) in the International System of Units (SI). During implementation, this constant can be directly invoked as a built-in parameter of the system, or stored and retrieved as a fixed entry in the chemical characteristic parameter library, ensuring the consistency of units in all calculations.

[0080] The temperature-corrected activation energy parameter, the pre-exponential factor, the gas constant, and the real-time temperature data extracted from the real-time environmental monitoring data are all input into the exponential transformation function. The exponential transformation function here transforms the thermodynamic and kinetic parameters of a chemical reaction, such as activation energy and temperature, into an exponential form that quantifies the tendency of the reaction to occur, using the mathematical form of the Arrhenius equation. This function is the core mathematical tool for converting activation energy into reaction rate or potential energy. In specific implementations, this function can be a predefined mathematical model or algorithm module. For example, in software programming, it can be encapsulated as a function that accepts activation energy, gas constant, and temperature as input parameters, internally executing the calculation logic of exp(-Ea / (R×T)). Alternatively, it can be implemented using a lookup table method, i.e., pre-calculating the exponential transformation results under different combinations of activation energy and temperature and storing them as a lookup table, then interpolating or directly searching based on real-time parameters during runtime.

[0081] The exponential transformation function performs a negative exponential operation on the ratio of the temperature-corrected activation energy parameter to the gas constant and the real-time temperature data, yielding an intermediate exponential value for each reagent pair. The negative exponential operation is a crucial step in the exponential transformation function, precisely quantifying the inhibitory effect of activation energy barriers on the tendency of a chemical reaction to occur. Higher activation energy results in a smaller negative exponent, indicating a more difficult reaction to occur. Conversely, lower activation energy results in a larger negative exponent, indicating a more likely reaction to occur. Furthermore, increasing temperature reduces the absolute value of the negative exponent, thus accelerating the reaction. In practice, this operation is typically performed using standard mathematical library functions, such as `math.exp()` in Python and `std::exp()` in C++. It is crucial to ensure that the units of the activation energy, gas constant, and real-time temperature data are matched and consistent before performing the operation to avoid calculation errors.

[0082] Multiplying the intermediate value of the exponential operation with the pre-exponential factor yields the reaction potential factor corresponding to each reagent pair. The multiplication operation combines the intermediate value of the exponential operation, obtained through negative exponentiation and reflecting the influence of the activation energy barrier, with the pre-exponential factor representing the molecular collision frequency. This combination is a complete embodiment of the Arrhenius equation, comprehensively considering both the frequency of effective collisions between reactant molecules and the possibility of overcoming the activation energy barrier, thus obtaining a comprehensive quantification of the reaction's tendency to occur. In practice, this is usually a direct multiplication operation; for example, in the calculation module, a simple product operation can be performed on two calculated values.

[0083] Through the above technical solution, this application can more accurately calculate the reaction potential factor for each reagent pair. Specifically, by retrieving the pre-exponential factor and gas constant from the chemical characteristic parameter library and inputting them, along with the temperature-corrected activation energy parameter and real-time temperature data, into the exponential transformation function, this application fully considers the influence of molecular collision frequency and real-time ambient temperature on the reaction rate. By performing a negative exponential operation on the ratio of activation energy to gas constant and real-time temperature data and multiplying it with the pre-exponential factor, this application can simulate the principle of the Arrhenius equation, converting the activation energy into a value related to the reaction rate constant, thereby accurately mapping the thermodynamic tendency of the chemical reaction. This overcomes the problem of inaccurate calculation of the reaction potential factor caused by ignoring molecular collision frequency and real-time temperature in traditional methods, making the subsequent generation of multi-reagent coupling risk matrices more accurate. This provides more reliable and refined basic data for risk assessment in the scenario of batch retrieval of multiple reagents in hazardous chemical cabinets, thereby improving the accuracy of overall risk perception and the effectiveness of access control strategies.

[0084] In some of the embodiments described above in this application, nonlinear coupling operations are proposed to calculate the base value of the reaction intensity for each pair of reagent combinations, thereby quantifying the comprehensive intensity of the chemical reaction in terms of thermodynamic drive and energy release dimensions. However, in its implementation, simple linear operations may not be able to accurately capture the complex synergistic amplification relationship between reaction potential energy and reaction heat, resulting in inaccurate calculation of the base value of intensity, which in turn affects the reliability of the subsequent coupling risk coefficient.

[0085] To address this, this application further proposes a nonlinear coupling operation between the reaction potential factor and the corresponding reaction heat to obtain the baseline reaction intensity value for each reagent pair. This process specifically includes: retrieving the nonlinear coupling coefficient corresponding to each reagent pair from a chemical characteristic parameter library; this nonlinear coupling coefficient is used to quantify the synergistic amplification effect between reaction potential and reaction heat; multiplying the reaction potential factor by the corresponding reaction heat to obtain a linear coupling intermediate value for each reagent pair; inputting this linear coupling intermediate value and the nonlinear coupling coefficient together into a power-law coupling function; using this power-law coupling function to fit and calculate the nonlinear amplification relationship between reaction potential and reaction heat, outputting the intensity amplification factor for each reagent pair; and multiplying this linear coupling intermediate value by the intensity amplification factor to obtain the baseline reaction intensity value for each reagent pair.

[0086] This nonlinear coupling coefficient is a key parameter for quantifying the synergistic amplification effect between reaction potential energy and reaction heat. Its role is to capture and characterize the complex nonlinear interactions that exist between thermodynamic driving forces (reaction potential energy) and energy release (reaction heat) during chemical reactions of different reagent combinations. These interactions are not simply linear superpositions, but may involve mutually promoting or inhibiting effects. For example, under certain conditions, some reactions may produce unexpectedly vigorous reactions even if the reaction potential energy and reaction heat are not high individually. This coefficient can be obtained in various ways. For instance, it can be extracted from the actual reaction behavior of different reagent combinations through regression analysis or machine learning training on large amounts of experimental data. Alternatively, it can be derived by predicting the energetic characteristics of intermolecular interactions at the microscopic level based on quantum chemical calculations or molecular dynamics simulations, combined with theoretical models. Another approach is to consult authoritative chemistry handbooks or databases for empirical nonlinear coupling parameters for specific reaction types or functional group combinations.

[0087] This intermediate value for linear coupling is the product of the reaction potential factor and the corresponding reaction heat. Its purpose is to provide a basic, linear reference value for subsequent nonlinear coupling calculations. During the calculation, the obtained reaction potential factor and the corresponding reaction heat are directly multiplied to obtain a preliminary value reflecting the linear superposition effect of the two. Although this value does not fully capture the nonlinear effect, it lays the foundation for subsequent nonlinear correction using a power-law coupling function.

[0088] This power-law coupling function is a mathematical model specifically designed to fit the nonlinear amplification relationship between reaction potential energy and reaction heat. Its function is to transform the intermediate value of linear coupling and the nonlinear coupling coefficient into an intensity amplification factor, thereby accurately reflecting the degree of nonlinear amplification effect. This function can take various forms; for example, it can be an exponential function of the form f(x, c) = x^c or f(x, c) = x × exp(c), where x represents the intermediate value of linear coupling and c represents the nonlinear coupling coefficient. The actual nonlinear amplification trend is fitted by adjusting the function parameters. Furthermore, more complex variants of the exponential function can be used, such as combinations containing multiple exponential or logarithmic terms. It can even learn and represent this complex nonlinear relationship by training machine learning models such as neural networks.

[0089] The intensity amplification factor is a multiplier calculated by the power-law coupling function based on the intermediate value of the linear coupling and the nonlinear coupling coefficient. Its function is to dynamically adjust the baseline reaction intensity value to quantify and reflect the degree of nonlinear amplification. Through this factor, the complex nonlinear relationship between reaction potential energy and reaction heat can be transformed into an operable numerical value. This value directly affects the intermediate value of the linear coupling, thus enabling the calculated baseline reaction intensity value to more accurately reflect the actual reaction intensity.

[0090] Through the above technical solution, this application effectively solves the problem that simple linear calculations cannot accurately capture the complex synergistic amplification relationship between reaction potential energy and reaction heat. Specifically, by retrieving the nonlinear coupling coefficient corresponding to each reagent pair from the chemical characteristic parameter library, the inherent characteristics of specific reagent combinations are accurately considered, avoiding errors that may be caused by simplification. Multiplying the reaction potential energy factor by the reaction heat yields an intermediate value for linear coupling, providing a solid foundation for subsequent nonlinear adjustments. More importantly, inputting this intermediate value for linear coupling and the nonlinear coupling coefficient together into a power-law coupling function allows for fitting calculations of the nonlinear amplification relationship between reaction potential energy and reaction heat. This accurately captures and quantifies complex synergistic amplification effects, solving the problem that traditional linear methods cannot accurately characterize. The output intensity amplification factor can dynamically adjust the amplification degree, making the calculated reaction intensity baseline value more accurately reflect the comprehensive intensity of the chemical reaction in both thermodynamic driving and energy release dimensions. This precise calculation of the intensity baseline provides a more reliable input for the subsequent generation of the multi-reagent coupled risk matrix, thereby improving the accuracy and reliability of risk assessment in the entire method of batch access and access control of multiple reagents in hazardous chemical cabinets. This makes the formulation of access control strategies more targeted and secure, and effectively prevents potential dangers caused by reagent incompatibility.

[0091] In some of the solutions mentioned above in this application, real-time concentration data and residual amount data are jointly mapped as dynamic weights for reaction probability to quantify the reaction probability. However, in this process, the synergistic effect between concentration and residual amount on the amplification of reaction probability is ignored, resulting in insufficient accuracy in risk calculation and failure to accurately reflect the synergistic contribution of actual reactant mass and contact probability.

[0092] To address this, this application further proposes a method for mapping real-time concentration data and remaining quantity data together as dynamic weights for reaction probabilities. This method includes: normalizing the real-time concentration data to obtain a concentration contribution coefficient corresponding to each reagent pair, which characterizes the degree to which the current concentration level promotes the reaction rate; normalizing the remaining quantity data by volume to obtain a mass contribution coefficient corresponding to each reagent pair, which characterizes the contribution of the total mass of reactants to the reaction scale; retrieving a synergistic amplification coefficient corresponding to each reagent pair from a chemical characteristic parameter library, which quantifies the amplification effect of the synergistic effect between concentration and remaining quantity on the reaction probability; and multiplying the concentration contribution coefficient, the mass contribution coefficient, and the synergistic amplification coefficient together to obtain the dynamic weights for the reaction probabilities of each reagent pair.

[0093] Specifically, the real-time concentration data refers to the actual concentration value of each reagent to be taken in the hazardous chemical cabinet at the current moment. This data reflects the number of molecules of reactants per unit volume and is a key factor affecting the chemical reaction rate. Methods for obtaining real-time concentration data may include, but are not limited to: real-time detection using optical sensors (e.g., spectrophotometers, refractometers) integrated into the reagent container; or measurement of solution properties using electrochemical sensors (e.g., conductivity sensors, pH sensors) and calculation using pre-defined calibration curves; or estimation using predictive models based on initial concentrations and historical consumption data. The remaining quantity data refers to the actual physical quantity of each reagent to be taken in the hazardous chemical cabinet at the current moment, typically expressed as volume or mass. This data determines the total amount of reactants that can participate in the reaction, thus affecting the potential scale and duration of the reaction. Methods for obtaining remaining quantity data may include, but are not limited to: real-time monitoring of reagent mass using weighing sensors (e.g., high-precision electronic scales) installed below the reagent container; or measuring the liquid level height using level sensors (e.g., ultrasonic level gauges, capacitive level gauges) and calculating the volume using the container geometry. Alternatively, the liquid output can be accumulated by a flow meter integrated at the outlet and deducted from the initial total.

[0094] Normalization of the real-time concentration data aims to unify concentration data of different dimensions and ranges onto a comparable scale, thereby obtaining the concentration contribution coefficient. This normalization process eliminates the influence of differences in the units and magnitudes of reagent concentrations, allowing the degree to which concentration promotes the reaction rate to be quantified fairly and accurately. For example, a min-max normalization method can be used to map the concentration value to the interval between 0 and 1, i.e., C_norm = (C - C_min) / (C_max - C_min). Alternatively, a Z-score normalization method can be used to convert the data into a distribution with a mean of 0 and a standard deviation of 1, i.e., C_norm = (C - μ) / σ. This concentration contribution coefficient, the result of normalization, is a dimensionless numerical value that directly characterizes the degree to which the current concentration level promotes the reaction rate; the larger the value, the stronger the promoting effect.

[0095] Volume normalization of the remaining amount data aims to unify remaining amount data of different dimensions and container sizes onto a comparable scale, thereby obtaining the mass contribution coefficient. This processing method ensures that the contribution of the remaining amount of different reagents to the reaction scale can be reasonably assessed, avoiding biases caused by differences in container size or reagent density. For example, the ratio of the current remaining amount to the initial total amount of reagents or the maximum capacity of the container can be calculated, i.e., V_norm = V_current / V_initial or V_norm = V_current / V_max. Alternatively, based on a preset risk level threshold, the remaining amount can be divided into different intervals and assigned corresponding contribution coefficients. This mass contribution coefficient is the result of volume normalization; it is a dimensionless numerical value used to characterize the degree of contribution of the total mass of substances that can participate in the reaction to the reaction scale. The larger the value, the larger the potential reaction scale.

[0096] The co-amplification factor is retrieved from the chemical characteristic parameter library. This library is a pre-built database that stores the physicochemical properties, reaction kinetic parameters, and specific parameters of their interactions for various chemical reagents. The co-amplification factor is a key parameter in this library, specifically used to quantify the nonlinear amplification effect on the reaction probability caused by the combined effects of concentration and remaining quantity. This factor is typically predetermined based on extensive experimental data, theoretical calculations, or expert experience, and is stored for specific reagent combinations. For example, for certain highly reactive reagent combinations, when both concentration and remaining quantity are high, the co-amplification factor may be much greater than 1, reflecting the exponential increase in risk. For less reactive reagents, the factor may be close to 1.

[0097] The dynamic weight of the reaction probability for each reagent pair is obtained by multiplying the concentration contribution coefficient, the mass contribution coefficient, and the synergistic amplification coefficient. This multiplication operation is a direct and effective aggregation method that organically combines the promoting effect of concentration on reaction rate, the contribution of the remaining amount to the reaction scale, and the synergistic amplification effect between the two into a comprehensive index. This dynamic weight of the reaction probability is a dimensionless value; the higher the value, the greater the probability and potential scale of the reaction of the reagent pair under the current real-time environment, and thus the higher its coupling risk.

[0098] Through the above technical solution, this application can more accurately quantify the reaction probability in multi-reagent batch retrieval scenarios. By normalizing real-time concentration data and remaining quantity data and introducing a synergistic amplification coefficient, this application solves the problem of neglecting the synergistic effect between concentration and remaining quantity in traditional methods. Specifically, the concentration contribution coefficient accurately reflects the degree to which the current concentration promotes the reaction rate, the mass contribution coefficient reasonably characterizes the contribution of the total mass of substances that can participate in the reaction to the reaction scale, and the synergistic amplification coefficient captures and quantifies the nonlinear amplification effect on the reaction probability when concentration and remaining quantity work together. Multiplying these three factors generates a more comprehensive and accurate dynamic weight for the reaction probability, thereby making the subsequent multi-reagent coupling risk calculation more accurate and able to more realistically reflect the chemical incompatibility risk of all reagent combinations in the batch retrieval list under real-time conditions, providing a more reliable data foundation for risk assessment and access control linkage of hazardous chemical cabinets.

[0099] In some embodiments of this application, a method is proposed to calculate the coupling risk coefficient based on the reaction intensity base value and the dynamic weight of the reaction probability and to aggregate and generate a multi-reagent coupling risk matrix. However, in its implementation, the catalytic or inhibitory effect of a third type of reagent on the reagent combination is ignored, resulting in inaccurate risk quantification.

[0100] To address this, this application further proposes a method for comprehensively calculating and matrix-aggregating coupling risk coefficients based on reaction intensity baselines, dynamic weights of reaction probabilities, and cross-coupling compensation mechanisms among multiple reagents, thereby generating a multi-reagent coupling risk matrix. This method includes: Based on the list of at least two reagents to be taken, an initial coupling risk matrix is ​​constructed, where each element of the initial coupling risk matrix corresponds to the initial position of a pair of reagent combinations.

[0101] Multiply the base value of the reaction intensity corresponding to each reagent pair by the dynamic weight of the reaction probability to obtain the basic coupling risk value of each reagent pair, and fill the corresponding matrix element position of the initial coupling risk matrix with the basic coupling risk value.

[0102] The cross-coupling compensation matrix corresponding to the at least two reagents to be used is retrieved from the chemical characteristic parameter library. Each compensation element of the cross-coupling compensation matrix is ​​used to quantify the catalytic or inhibitory effect of the third type of reagent on the corresponding reagent combination.

[0103] The initial coupling risk matrix after filling is multiplied element-wise with the cross-coupling compensation matrix, and the basic coupling risk value of each reagent pair is amplified by catalytic effect or attenuated by inhibitory effect to obtain the multi-reagent coupling risk matrix.

[0104] Specifically, when constructing the initial coupling risk matrix, the system obtains a list of at least two reagents to be taken from the user's batch retrieval request. This retrieval list clarifies all reagent types involved in the operation. Based on this retrieval list, the system constructs an initial coupling risk matrix. This matrix is ​​typically an N x N two-dimensional data structure, where N is the number of reagents in the retrieval list. The rows and columns of the matrix correspond to different reagents in the list, and each matrix element (i.e., a cell) is reserved to store the potential risk value between a pair of reagent combinations. For example, it can be represented as a two-dimensional array or an adjacency matrix, where the diagonal elements can be set to zero or represent the stability risk of the reagent itself, while the off-diagonal elements are used to represent the coupling risk between different reagents. This structured construction provides a clear framework for subsequent risk calculation and aggregation, ensuring that the risks of all pairwise reagent combinations are systematically considered and recorded.

[0105] Based on this, to quantify the basic risk level of each reagent pair, the base value of the reaction intensity corresponding to each reagent pair is multiplied by the dynamic weight of the reaction probability to obtain the basic coupling risk value of each reagent pair. The base value of the reaction intensity, according to the above implementation method, quantifies the comprehensive intensity of the chemical reaction in both thermodynamic driving and energy release dimensions, reflecting the inherent danger of the chemical reaction itself. The dynamic weight of the reaction probability, according to the above implementation method, characterizes the synergistic contribution of the actual reactant mass and the reaction contact probability to the risk level, reflecting the probability of the reaction occurring in a real-time environment. Multiplying these two values ​​allows for a comprehensive consideration of the inherent danger of the reaction and the actual probability of its occurrence, resulting in a more comprehensive basic coupling risk value. These calculated basic coupling risk values ​​are then accurately filled into the reserved positions corresponding to the reagent pairs in the initial coupling risk matrix. For example, if the i-th row and j-th column of the matrix represents the combination risk of reagent i and reagent j, then the basic coupling risk values ​​of reagent i and reagent j are filled into that position.

[0106] To further improve the accuracy of risk assessment, this application introduces a cross-coupling compensation mechanism. Specifically, a cross-coupling compensation matrix corresponding to the at least two reagents to be used is retrieved from a pre-stored chemical characteristic parameter library. This chemical characteristic parameter library is a pre-established database that stores a large number of chemical substances' physicochemical properties, reaction characteristics, safety data, and various reaction parameters, including the cross-coupling compensation matrix required in this step. The cross-coupling compensation matrix is ​​a matrix specifically used to quantify the impact of a "third type of reagent" in a multi-reagent system on the reaction risk between specific pairwise reagent combinations. Each compensation element is a correction factor used to adjust the base coupling risk value. For example, when reagents A, B, and C are used, the compensation matrix will include the effect of C on the AB combination, the effect of B on the AC combination, etc. The compensation element is a specific value in the cross-coupling compensation matrix. When the compensation element value is greater than 1, it indicates that the third type of reagent has a catalytic effect on the current reagent combination, amplifying its risk. When the compensation element value is less than 1, it indicates that the third type of reagent has an inhibitory effect, attenuating its risk. These values ​​can be pre-stored in the chemical characteristic parameter library based on experimental data, theoretical calculations, or expert experience. Category 3 reagents refer to any other reagents in the bulk purchase list besides the two reagents currently being assessed for risk.

[0107] The initial coupling risk matrix, populated with basic coupling risk values, is multiplied element-wise with the cross-coupling compensation matrix. This multiplication amplifies the catalytic effect or attenuates the inhibitory effect on the basic coupling risk value for each reagent pair, resulting in a multi-reagent coupling risk matrix. Element-wise multiplication involves multiplying corresponding elements of the two matrices. By multiplying with the compensation element, the basic coupling risk value is dynamically adjusted based on the influence of the third type of reagent. If the compensation element is greater than 1, the basic risk value is amplified, reflecting the increased risk due to catalysis. If the compensation element is less than 1, the basic risk value is attenuated, reflecting the reduced risk due to inhibition. This corrected multi-reagent coupling risk matrix comprehensively and accurately quantifies the chemical incompatibility risk of all reagent combinations within the batch pick-up list under real-time conditions, and considers the complex interactions in multi-reagent systems.

[0108] Through the above technical solution, this application overcomes the limitation of traditional risk assessment methods that neglect the catalytic or inhibitory effects of a third-class reagent on a specific reagent combination in a multi-reagent system. Specifically, by constructing an initial coupling risk matrix and filling it with basic coupling risk values, a structured foundation for risk assessment is provided. Based on this, a cross-coupling compensation matrix is ​​retrieved from a chemical characteristic parameter library and applied. The compensation elements of this matrix can accurately quantify the catalytic or inhibitory effects of a third-class reagent on the target reagent combination. By multiplying the basic coupling risk values ​​element-wise with the compensation elements, the system can dynamically amplify or attenuate the basic risk, thereby generating a more comprehensive, accurate, and dynamic multi-reagent coupling risk matrix. This makes the quantitative characterization of the chemical incompatibility risk of all reagent combinations in the batch access list more refined and realistic in a real-time environment, improving the accuracy of risk perception and providing a more reliable basis for subsequent risk-level-based access control strategies, effectively avoiding potential security risks caused by complex interactions among multiple reagents.

[0109] In some of the embodiments described above in this application, a multi-reagent coupling risk matrix is ​​proposed to be generated based on the reaction intensity base value, the dynamic weight of the reaction probability, and the cross-coupling compensation mechanism. However, in its implementation, there is a lack of detailed description on how to specifically apply the compensation mechanism to amplify or attenuate the risk value, which may result in an inaccurate risk assessment and fail to fully reflect the true impact of the third type of reagent.

[0110] To address this, this application further proposes to amplify the catalytic effect or attenuate the inhibitory effect on the basic coupling risk value of each reagent pair to obtain a multi-reagent coupling risk matrix, specifically including: The compensation element value corresponding to each reagent combination is extracted from the cross-coupling compensation matrix. When the compensation element value is greater than 1, it is used to characterize the catalytic effect of the third type of reagent on the current reagent combination. When the compensation element value is less than 1, it is used to characterize the inhibitory effect of the third type of reagent on the current reagent combination.

[0111] Multiply each basic coupling risk value in the filled initial coupling risk matrix by the corresponding compensation element value to obtain the compensated risk value for each reagent pair.

[0112] The compensated risk values ​​of all reagent combinations are filled into the matrix structure according to the correspondence of the reagent combinations to obtain the multi-reagent coupling risk matrix.

[0113] Specifically, compensation element values ​​corresponding to each reagent pair are extracted from the cross-coupling compensation matrix. These compensation element values ​​are key factors in quantifying the influence of the third-class reagent on a specific reagent combination, directly determining the direction and magnitude of risk value adjustment. These values ​​can be pre-calculated based on extensive chemical experimental data, quantum chemical calculations, or molecular dynamics simulations and stored in a chemical feature parameter library. These values ​​reflect changes in the activation energy or reaction pathway of the target reagent combination reaction under specific environmental conditions in the presence of the third-class reagent. Furthermore, machine learning models (such as neural networks or support vector machines) can be trained, inputting features such as the molecular structure, functional groups, and electronic properties of the reagents, to output predicted compensation element values. For reagent combinations with known catalytic or inhibitory effects, an expert rule base can be established, allowing direct querying of corresponding compensation element values ​​based on reagent category and functional group matching rules.

[0114] The compensated risk value for each reagent combination is obtained by multiplying each basic coupling risk value in the filled initial coupling risk matrix with its corresponding compensation element value. This step is the core operation for dynamically adjusting the risk value. Through multiplication, the compensation element value is directly applied to the basic coupling risk value, thereby amplifying or attenuating the risk. In the software implementation, the initial coupling risk matrix and the diagonal matrix composed of compensation element values ​​(or the compensation matrix with the same dimension as the initial matrix) can be multiplied element-wise (Hadamard product) to efficiently adjust the risk value of all reagent combinations. Alternatively, for each reagent combination, the system retrieves the corresponding compensation element value from the cross-coupling compensation matrix, then multiplies this value with the basic coupling risk value of that reagent combination to calculate the compensated risk value one by one.

[0115] The compensated risk values ​​for all reagent combinations are filled into a matrix structure according to the correspondence between the reagent combinations, resulting in the multi-reagent coupling risk matrix. This step aims to ensure that the compensated risk data can be stored and subsequently analyzed in a structured and easily processed form (matrix). In programming, a two-dimensional array or nested list can be used to represent the matrix structure, and the calculated compensated risk values ​​are filled into the corresponding positions in the matrix according to their corresponding reagent combination indices (e.g., row index represents reagent A, column index represents reagent B). If the number of reagent combinations is large and most combinations do not have direct coupling risks (i.e., the compensated element value is 1 or the basic risk value is 0), sparse matrix storage techniques (such as coordinate lists, compressed sparse rows / columns, etc.) can be used to save storage space and improve processing efficiency.

[0116] Through the above technical solution, this application can accurately quantify the real impact of the third type of reagent on a specific reagent combination, overcoming the limitations of the ambiguous application of compensation mechanisms in traditional methods. Specifically, by extracting compensation element values ​​from the cross-coupling compensation matrix and clarifying that a value greater than 1 indicates a catalytic effect and a value less than 1 indicates an inhibitory effect, this solution achieves dynamic and adaptive adjustment of risk values. When a catalyst is present, the risk value is amplified, more accurately reflecting the increase in potential danger. When an inhibitor is present, the risk value is attenuated, avoiding overly conservative risk assessment. This risk value adjustment based on actual chemical interactions enables the generated multi-reagent coupling risk matrix to more realistically and accurately characterize the chemical incompatibility risk of all reagent combinations in the batch access list under real-time conditions, providing more reliable input for subsequent spatiotemporal risk field evolution models. In addition, filling the matrix structure with the compensated risk values ​​maintains the organization and readability of the risk data, facilitating efficient risk analysis, visualization, and the formulation of access control strategies. Furthermore, based on generating basic coupling risk values ​​and filling the initial matrix, this solution introduces a refined cross-coupling compensation mechanism, making the construction of the multi-reagent coupling risk matrix more complete and accurate. It not only considers the basic risks between any two reagents, but also takes into account the catalytic or inhibitory effects of other reagents in the environment (Class III reagents) on these paired risks. This allows the generated risk matrix to more comprehensively and accurately reflect the real risk situation in complex multi-reagent environments, improving the risk perception accuracy and safety of the method for batch access and access control of multiple reagents in hazardous chemical cabinets.

[0117] In some of the solutions mentioned above in this application, a multi-reagent coupled risk matrix and the spatial location data inside the cabinet are input into a spatiotemporal risk field evolution model to output a four-dimensional spatiotemporal risk field, which is used to quantify the distribution and evolution trend of risk in the spatial and temporal dimensions. However, in this process, how to specifically realize the dynamic iterative solution of the risk field to accurately simulate the spatial propagation and temporal decay of risk, as well as the updating of risk distribution by combining new data, has not been explained in detail, resulting in insufficient real-time performance and accuracy of risk prediction.

[0118] In response, this application further proposes a method for batch access and access control of multiple reagents in hazardous chemical cabinets. The method involves inputting the multi-reagent coupled risk matrix and the cabinet's spatial location data obtained from the sensor network into a spatiotemporal risk field evolution model. This model iteratively solves the distribution and evolution trend of risks in the spatial and temporal dimensions, and outputs a four-dimensional spatiotemporal risk field under the current operating scenario.

[0119] Specifically, see Figure 5 The method includes the following steps.

[0120] 501. Use the multi-reagent coupling risk matrix as the risk source term, and input it together with the spatial location data inside the cabinet into the spatiotemporal risk field evolution model to initialize the initial risk distribution field of the spatiotemporal risk field evolution model at the current moment.

[0121] The multi-reagent coupling risk matrix quantifies the chemical incompatibility risk of all reagent combinations within the batch retrieval list in a real-time environment. As a risk source term in the spatiotemporal risk field evolution model, it provides initial intensity and distribution information of the risk, serving as the starting point for risk evolution calculations. This matrix can be an N×N symmetric matrix, where N is the number of reagents to be retrieved, and matrix elements represent the coupling risk value between any two reagents. The cabinet spatial location data describes the physical layout information of each storage location within the hazardous materials cabinet, such as the coordinates, size, adjacency relationships, and the types and quantities of reagents stored in each storage cell. This data can be acquired in real-time by sensor networks (such as RFID positioning systems and visual recognition systems) or by pre-defined cabinet structure data. For example, a three-dimensional coordinate system can be used to accurately represent each storage location, or a topological graph can be used to represent the adjacency relationships between storage locations. This spatiotemporal risk field evolution model aims to simulate the dynamic distribution and evolution trend of risks within the hazardous materials cabinet in both spatial and temporal dimensions. It comprehensively considers the generation, propagation, attenuation, and updating of risks. This model can be constructed based on partial differential equations (such as reaction-diffusion equations) and the dynamic changes of the risk field can be solved numerically. Alternatively, an agent-based simulation method can be used to simulate the diffusion and interaction of risk sources in space. Cellular automata models can also be employed, dividing the cabinet space into discrete units and simulating the propagation and evolution of risk by defining local rules. This initial risk distribution field is a snapshot of the risk state at the current moment in the spatiotemporal risk field evolution model, reflecting the initial distribution of the immediate risk brought about by the current bulk retrieval request in the cabinet space. It is typically initialized by a multi-reagent coupled risk matrix and cabinet spatial location data; for example, the risk values ​​in the coupled risk matrix are mapped to the spatial locations of the corresponding reagents, and their initial diffusion range is considered.

[0122] 502. The initial risk distribution field is calculated by performing neighborhood diffusion calculation using the spatial propagation operator in the spatiotemporal risk field evolution model to obtain the spatial propagation components of the risk between different spatial locations within the cabinet.

[0123] The spatial propagation operator describes the diffusion and spread mechanism of risk between different spatial locations within a hazardous materials container. Its function is to calculate the degree to which risk propagates from one location to its neighboring locations. For example, risk propagation weights can be calculated based on a distance decay function, with greater weights for closer locations. Alternatively, physical barriers (such as partitions) can be used to restrict risk propagation, allowing propagation only in directions without barriers or with weaker barriers. The neighborhood diffusion calculation is the specific computational process performed by the spatial propagation operator. By considering the influence of the risk source on its surrounding neighborhood, it simulates the spatial diffusion of risk. For example, convolution operations can be used to convolve the initial risk distribution field with a diffusion kernel function to simulate the smooth diffusion of risk. Alternatively, a graph-based method can be used, treating the storage locations within the container as nodes in a graph, and risk propagation as a flow on the graph. The diffusion path and intensity are determined by calculating the connection strength between nodes. This spatial propagation component represents the distribution change of risk within the container space due to the diffusion effect. It quantifies the trend and intensity of risk spreading from high-risk areas to low-risk areas and is an important component of the dynamic changes in the risk field in the spatial dimension.

[0124] 503. The initial risk distribution field is subjected to time-series evolution calculation by using the time decay operator in the spatiotemporal risk field evolution model to obtain the natural decay component of risk over time.

[0125] The time decay operator describes the mechanism by which risk naturally diminishes over time. Its function is to calculate the degree to which the risk value decreases over time without new risk sources entering or spreading. For example, an exponential decay function can be used, where the risk value decreases exponentially over time. Alternatively, a linear decay model can be used, where the risk value decreases at a constant rate over time. The time-series evolution calculation is the specific calculation process performed by the time decay operator. By considering the influence of time factors on the risk value, it simulates the dynamic changes of risk over time. For example, iterative calculations can be performed using discrete time steps, applying the decay function within each time step. Alternatively, a continuous-time model can be used, solving differential equations to describe the continuous changes of risk over time. This natural decay component represents the decline of risk over time due to its inherent characteristics or the self-cleaning capacity of the environment. It quantifies the trend and intensity of risk weakening over time and is an important component of the dynamic changes of the risk field over time.

[0126] 504. Based on the spatial propagation component, the time decay component, and the newly added reagent coupling risk matrix input at the next moment, perform iterative fusion calculation to output the four-dimensional spatiotemporal risk field under the current operation scenario.

[0127] The newly added multi-reagent coupling risk matrix is ​​new coupling risk information generated in subsequent time steps due to new batch retrieval requests or environmental changes. It serves as a new risk source input to update the spatiotemporal risk field, ensuring the model reflects the latest operational scenarios and risk situations. This iterative fusion calculation integrates multiple factors, including spatial propagation components, temporal decay components, and the newly added multi-reagent coupling risk matrix, to update and evolve the spatiotemporal risk field. For example, a weighted summation method can be used to superimpose the components according to their importance. Alternatively, a more complex numerical integration method can be used to solve the dynamic equations of the risk field within each time step. This four-dimensional spatiotemporal risk field is the output of the spatiotemporal risk field evolution model. It is risk distribution data containing four dimensions: spatial (X, Y, Z coordinates) and temporal (T). This risk field can dynamically display the risk intensity at any location within the hazardous materials container at any time, providing comprehensive and detailed risk information for subsequent access control strategies.

[0128] This application effectively addresses the shortcomings in real-time performance and accuracy of risk prediction in related technologies by introducing a spatiotemporal risk field evolution model and detailing its iterative solution mechanism. Specifically, by inputting the multi-reagent coupled risk matrix as a risk source term along with the spatial location data within the cabinet into the model, the initial risk distribution field at the current moment can be accurately initialized, laying a solid foundation for subsequent dynamic evolution. Based on this, a spatial propagation operator performs neighborhood diffusion calculations on the initial risk distribution field, generating a spatial propagation component. This allows for accurate simulation of the risk propagation process, avoiding the limitation of only considering point source risks while ignoring their diffusion effects. Simultaneously, a time decay operator performs temporal evolution calculations on the initial risk distribution field, generating a natural decay component. This effectively captures the natural decay characteristics of risk over time, preventing unreasonable accumulation of risk values ​​and making the risk assessment results more realistic. By iteratively fusing the spatial propagation component, the time decay component, and the multi-reagent coupled risk matrix added at the next moment, this application can output the four-dimensional spatiotemporal risk field under the current operating scenario in real time and dynamically. This iterative fusion mechanism ensures that the risk field can be continuously updated, reflecting not only the current risk situation but also predicting future evolution trends. This greatly improves the real-time nature and accuracy of risk perception, providing a more refined and reliable risk basis for the dynamic access control of hazardous chemical storage containers.

[0129] In some of the above-mentioned schemes in this application, a neighborhood diffusion calculation is proposed to calculate the propagation component of risk in the spatial dimension by performing neighborhood diffusion calculation on the initial risk distribution field through a spatial propagation operator. However, in this process, due to the lack of specific methods for determining the neighborhood location, the method for obtaining the diffusion weight, and detailed steps for risk value fusion calculation, the calculation of the risk propagation model may be inaccurate, affecting the accuracy of the risk field evolution.

[0130] To address this, this application further proposes using the spatial propagation operator in the spatiotemporal risk field evolution model to perform neighborhood diffusion calculations on the initial risk distribution field, obtaining the spatial propagation components of risk between different spatial locations within the cabinet. Specifically, this process includes: based on the spatial location data within the cabinet, determining the neighborhood location set for each storage location within the hazardous materials cabinet. This neighborhood location set includes other storage locations physically adjacent to the current storage location. This step aims to clarify the spatial relationships between various storage units within the hazardous materials cabinet, laying the foundation for subsequent spatial propagation calculations of risk. Determining the neighborhood location set is crucial for simulating risk diffusion, as it defines the adjacent areas that the risk may directly affect. For example, each storage location can be coordinate-encoded using a pre-established 3D model or 2D plan layout of the cabinet. Once a storage location is determined, its Euclidean or Manhattan distance to all other storage locations is calculated, and a distance threshold is set; storage locations with distances less than this threshold are identified as neighborhood locations. Alternatively, a list of physically adjacent storage location IDs can be pre-defined in the metadata of each storage location during the hazardous materials cabinet design phase. For example, in a grid-like storage structure, each cell can predefine its adjacent cells (up, down, left, right, and front, if it is a three-dimensional structure) as its neighborhood.

[0131] The diffusion coefficient set corresponding to the spatial propagation operator is retrieved from the spatiotemporal risk field evolution model. This set of diffusion coefficients contains the risk diffusion weight between the current storage location and each neighboring location in the set of neighboring locations. The diffusion coefficient set is a key parameter for quantifying the intensity of risk propagation between different storage locations. The risk diffusion weight reflects the ease or magnitude of risk propagation from one location to its neighboring locations, and it may be affected by factors such as physical isolation between storage locations, ventilation conditions, and reagent types. For example, the diffusion coefficient set can be pre-calibrated and stored through expert experience assessment, historical accident data analysis, or methods based on fluid dynamics / chemical reaction diffusion simulation. Different diffusion weights can be assigned to partitions of different materials and different ventilation opening designs. Furthermore, the diffusion coefficient set can also be dynamically adjusted; for example, based on real-time environmental monitoring data (such as ventilation system operating status and partition integrity sensor data), the risk diffusion weight between different neighborhoods can be calculated or corrected in real time using a machine learning model.

[0132] The system extracts the current risk value of the current storage location and the neighborhood risk value of each neighboring location in the set of neighboring locations from the initial risk distribution field. The initial risk distribution field is a snapshot of the risk status of all storage locations within the hazardous materials container at the current moment, taken by the spatiotemporal risk field evolution model. Extracting the current risk value and neighborhood risk values ​​is a direct input for spatial propagation calculations, ensuring that the calculations are based on the latest risk status. For example, the initial risk distribution field can be a multidimensional array or matrix, where each element corresponds to the risk value of a storage location. The system directly retrieves the risk value of a specified storage location and its neighborhood from this data structure through indexing or query operations. Alternatively, the risk values ​​can be stored in a distributed database, with the risk value of each storage location as a record. The system retrieves the risk values ​​of the current location and its neighborhood in batches based on the storage location ID through a query interface.

[0133] The spatial propagation component of the current storage location is obtained by weighted summation of the neighborhood risk value and its corresponding risk diffusion weight, and then fusion calculation with the current risk value. This step is the core calculation process of the spatial propagation operator, aiming to comprehensively consider the impact of neighborhood risk on the current location and the risk state of the current location itself, thereby quantifying the propagation effect of risk in the spatial dimension. The weighted summation reflects the differences in the impact of different neighborhoods on the current location, while the fusion calculation combines this external influence with the internal state. For example, the weighted summation can adopt a simple linear weighted model, i.e., ∑(neighborhood risk value × risk diffusion weight). The fusion calculation can linearly superimpose the weighted summation result with the current risk value, for example, spatial propagation component = current risk value + α × (weighted summation result), where α is the fusion coefficient. In addition, the fusion calculation can also adopt more complex nonlinear models, such as fusion algorithms based on neural networks or fuzzy logic, to better capture the complex nonlinear characteristics of risk propagation.

[0134] Through the above technical solution, this application overcomes the problem of inaccurate risk propagation model calculations in related technologies. By defining and implementing the fusion calculation of neighborhood location, diffusion weight, and risk value in detail, the spatiotemporal risk field evolution model makes the risk propagation calculation in the spatial dimension more accurate and reliable. This not only improves the real-time perception capability of risk distribution within hazardous chemical storage containers but also provides a more solid and refined data foundation for subsequent access control strategies based on four-dimensional spatiotemporal risk fields, thereby effectively enhancing the safety management level of hazardous chemical storage containers and reducing potential accident risks.

[0135] In some of the solutions mentioned above in this application, a time decay operator is proposed to perform time-series evolution calculations on the initial risk distribution field to quantify the natural decay of risk over time. However, in this process, the time decay calculation may not clearly define how to accurately determine the time step and decay coefficient based on real-time environmental data, or how to efficiently perform exponential decay calculations, resulting in inaccurate and unreliable calculations of the natural decay component of risk.

[0136] To address this, this application further proposes a method to perform temporal evolution calculations on the initial risk distribution field using a time decay operator in a spatiotemporal risk field evolution model, thereby obtaining the natural decay component of risk over time. This method includes: extracting timestamp information from real-time environmental monitoring data for the current operational scenario, and determining the time step between the current iteration and the previous iteration based on this timestamp information; retrieving the time decay coefficient corresponding to the time decay operator from the spatiotemporal risk field evolution model, whereby the time decay coefficient quantifies the natural decay rate of the risk value per unit time; inputting the risk value, the time step, and the time decay coefficient at each location in the initial risk distribution field into the time decay operator, and performing exponential decay calculations on the risk value at each location using the time decay operator to obtain the natural decay component at each location.

[0137] Specifically, when extracting timestamp information for the current operational scenario from real-time environmental monitoring data and determining the time step between the current iteration and the previous iteration based on this timestamp information, the system continuously collects real-time environmental monitoring data within the hazardous materials container via a sensor network. This data typically contains precise timestamp information. During iterative calculations of the spatiotemporal risk field evolution model, the system acquires the timestamp at the start of the current calculation cycle and the timestamp at the end of the previous calculation cycle, using the difference between the two to precisely determine the time step between the current iteration and the previous iteration. This dynamic determination of the time step ensures the real-time performance and accuracy of risk evolution calculations, effectively avoiding error accumulation that might result from using a fixed time step. It is particularly effective in adaptively adjusting the computational granularity when data acquisition frequency or system processing capacity changes. Furthermore, the time step can also be adaptively determined by analyzing the timestamp sequence in the real-time environmental monitoring data stream to detect the frequency or interval of data updates. For example, when data updates are frequent, the time step can be smaller to capture more refined risk evolution. When data updates are infrequent, the time step can be larger to reduce computational burden.

[0138] When retrieving the time decay coefficient corresponding to the time decay operator from the spatiotemporal risk field evolution model, this coefficient is a key parameter in the model used to quantify the natural decay rate of the risk value per unit time. This coefficient can be pre-calibrated using extensive experimental data, historical accident analysis, chemical property assessment, or expert experience, and stored in the model's parameter library. For example, for volatile, easily degradable, or easily diluted chemicals, the risk decay coefficient will be relatively large, indicating a faster risk decay rate. Conversely, for highly stable chemicals that are not easily volatile or degradable, the decay coefficient will be relatively small. In some implementations, the time decay coefficient can also be dynamically adjusted, for example, by using a machine learning model to calculate or correct in real time based on factors such as ventilation conditions, reagent volatility, and environmental purification capacity from real-time environmental monitoring data, to more accurately reflect the risk decay pattern in the actual environment.

[0139] When the risk value, time step, and time decay coefficient at each location in the initial risk distribution field are input into the time decay operator, and the time decay operator performs exponential decay calculation on the risk value at each location to obtain the natural decay component of each location, the exponential decay calculation is the core function of the time decay operator, used to simulate the natural decay process of the risk value over time. In specific implementation, the system traverses each storage location in the initial risk distribution field to obtain the current risk value at that location. Then, the risk value, the determined time step, and the corresponding time decay coefficient are input into the preset exponential decay function. This function is usually in the form R_new = R_old × exp(-k × dt), where R_old is the initial risk value, R_new is the decayed risk value, k is the time decay coefficient, and dt is the time step. Through this calculation, the natural decay amount of each location within the current time step, i.e., R_old - R_new, can be obtained, thus accurately reflecting the natural decay of the risk. In addition to the exponential decay model, other nonlinear decay models, such as power-law decay or logarithmic decay, can also be used. The specific choice depends on the fitting effect on the decay law of different types of risks and the needs of the actual application scenario.

[0140] Through the above technical solution, this application can accurately quantify the natural decay component of risk over time. By extracting timestamp information from real-time environmental monitoring data and determining the time step accordingly, the real-time nature and accuracy of the time interval are ensured, effectively avoiding the error accumulation that may be caused by a fixed time step, making the risk evolution calculation closer to reality. By retrieving preset or dynamically adjusted time decay coefficients from the spatiotemporal risk field evolution model, a scientific and objective basis is provided for risk decay calculation, enhancing the reliability of the calculation. By inputting the risk value, time step, and time decay coefficient of each location in the initial risk distribution field into the time decay operator for exponential decay calculation, this model conforms to the physical law of natural risk decay, achieving refined processing of the risk value at each location. This enables the spatiotemporal risk field evolution model to more accurately predict the distribution and evolution trend of risk in the time dimension, providing a more reliable risk assessment basis for subsequent access control strategies, thereby improving the accuracy and safety of the overall risk management of hazardous chemical containers.

[0141] In some of the embodiments described above in this application, an iterative fusion calculation based on spatial propagation components and time decay components is proposed to output a four-dimensional spatiotemporal risk field. However, in its implementation, there are shortcomings in how to accurately fuse spatial propagation components, time decay components, and newly added reagents coupled risk matrices to ensure the accuracy and timeliness of risk field updates, thereby avoiding risk prediction deviations and dynamic access control failures.

[0142] To address this, this application further proposes a method for iteratively fusing and calculating the four-dimensional spatiotemporal risk field under the current operational scenario based on the aforementioned spatial propagation component, the aforementioned time decay component, and the newly added multi-reagent coupling risk matrix input at the next moment. Specifically, this method includes: adding the initial risk distribution field at the current moment to the spatial propagation component to obtain the intermediate risk distribution field after spatial propagation; subtracting the intermediate risk distribution field after spatial propagation from the time decay component to obtain the intermediate risk distribution field after time decay; and adding the intermediate risk distribution field after time decay to the newly added multi-reagent coupling risk matrix to obtain the updated risk distribution field at the next moment. This updated risk distribution field is then used as the four-dimensional spatiotemporal risk field under the current operational scenario.

[0143] The process involves adding the initial risk distribution field at the current moment to the spatial propagation component to obtain the intermediate risk distribution field after spatial propagation. This aims to integrate the current risk baseline with the spatial diffusion effect of risk. The initial risk distribution field represents the risk state within the hazardous materials container at the start of the current iteration; it may have evolved from the risk at the previous moment or been set by initial conditions. The spatial propagation component quantifies the degree to which risk spreads from one storage location to its neighboring locations. By superimposing these two components, the spread of risk in physical space can be accurately simulated, resulting in an intermediate risk state that reflects the impact of spatial diffusion. This can be achieved, for example, by adding each risk value in the initial risk distribution field to its corresponding spatial propagation component element-wise, or by using a weighted average, where the spatial propagation component is assigned different weights based on its contribution to the risk at the current location.

[0144] Furthermore, subtracting the time decay component from the intermediate risk distribution field after spatial propagation yields the intermediate risk distribution field after time decay, aiming to simulate the natural decline of risk over time. The intermediate risk distribution field after spatial propagation already includes the spatial diffusion effect of risk, while the time decay component characterizes the tendency of risk to decrease naturally over time without new risk sources or external intervention, such as due to the completion of chemical reactions, the dissipation of volatiles, or the effect of ventilation systems. By subtracting the time decay component from the intermediate risk distribution field after spatial propagation, unreasonable overestimation of risk values ​​can be avoided, allowing the risk field to more realistically reflect the risk level at the current moment. Specifically, the corresponding time decay component can be subtracted from each risk value in the intermediate risk distribution field after spatial propagation, or the exponential decay of risk can be simulated by multiplying by a time decay factor less than 1.

[0145] Based on this, the intermediate risk distribution field after time decay is added to the newly added multi-reagent coupled risk matrix to obtain the updated risk distribution field at the next moment. This updated risk distribution field is then used as the four-dimensional spatiotemporal risk field under the current operational scenario. The intermediate risk distribution field after time decay has already comprehensively considered the spatial diffusion and time decay effects of risks, while the newly added multi-reagent coupled risk matrix represents the potential risks that may be introduced due to new batch retrieval requests or reagent operations under the current operational scenario. By superimposing these two, new risk sources can be incorporated into the calculation of the risk field in a timely manner, thereby generating an updated risk distribution field that comprehensively reflects all known risk factors under the current operational scenario. This updated risk distribution field, combining the effects of space, time, and new events, constitutes the four-dimensional spatiotemporal risk field under the current operational scenario, providing a real-time and accurate data foundation for subsequent risk assessment and access control. For example, the intermediate risk distribution field after time decay can be added element-wise to the newly added multi-reagent coupled risk matrix, or more complex fusion algorithms, such as Bayesian update or Kalman filtering, can be used to more accurately integrate old and new risk information.

[0146] Through the above technical solution, this application can accurately integrate the spatial propagation, temporal decay, and new risk input of risks, effectively solving the problem of inaccurate risk field updates. Specifically, adding the initial risk distribution field to the spatial propagation component can accurately simulate the diffusion process of risks within the cabinet space, avoiding underestimation of risks in adjacent areas. By subtracting the temporal decay component from the intermediate risk distribution field after spatial propagation, the characteristics of risk naturally fading over time are effectively simulated, preventing unreasonable overestimation of risk values ​​and ensuring the authenticity and timeliness of the risk field. Furthermore, adding the intermediate risk distribution field after time decay to the newly added multi-reagent coupled risk matrix can promptly incorporate new potential risk sources into the risk field calculation, ensuring that the risk field always reflects the latest operational scenarios and potential hazards. This iterative fusion calculation mechanism enables the generated four-dimensional spatiotemporal risk field to reflect the risk distribution and evolution trend within the hazardous chemical cabinet in real time and dynamically, providing a highly accurate and reliable basis for subsequent risk level assessment and dynamic access control, improving the safety and response efficiency of hazardous chemical management, and avoiding potential safety hazards caused by risk prediction deviations and dynamic access control failures.

[0147] In some of the above-mentioned solutions in this application, an access control strategy is proposed based on the comparison results of the real-time field value of the four-dimensional spatiotemporal risk field and the multi-level security thresholds to dynamically adjust permissions to deal with risks. However, in its implementation, relying solely on the comparison of the current field value may ignore the gradient changes of risks in space and the evolution trend in future time, resulting in inaccurate permission adjustments and an inability to respond in a timely manner to the rapid increase of the risk diffusion frontier region or to predict future risk escalation.

[0148] In response, this application further proposes a method for batch access and access control of multiple reagents in hazardous chemical cabinets. When comparing the real-time field value of the four-dimensional spatiotemporal risk field with multi-level safety thresholds to obtain and execute an access control strategy bound to the current risk level, the method includes the following steps: Spatial gradient analysis is performed on the real-time field value of the four-dimensional spatiotemporal risk field to obtain the risk field change gradient between different storage locations within the cabinet. Based on the risk field change gradient, at least one risk diffusion front region where the risk field strength increases rapidly is identified.

[0149] Based on the time-series variation data of the real-time field value within at least one risk diffusion frontier region, a risk evolution rate corresponding to each risk diffusion frontier region is fitted and generated. This risk evolution rate is used to characterize the speed and intensity growth trend of risk spread in the future.

[0150] The real-time field value, the risk evolution rate, and the multi-level security thresholds are input into the permission policy dynamic generation model. The dynamic generation model of permission policy comprehensively calculates the current risk status and future risk trends, outputs the permission control policy bound to the current risk level, and executes it.

[0151] The spatial gradient analysis of the real-time field value of this four-dimensional spatiotemporal risk field aims to quantify the rate and direction of risk change in space, thereby identifying areas where risk is rapidly growing or spreading. This is crucial for accurately locating potential hazards and predicting risk spread paths. Several methods can be employed for implementation. For example, the local gradient can be determined by calculating the difference in risk field values ​​between adjacent storage locations and combining this with the physical distance between these locations. For instance, using the finite difference method or central difference method, the risk field value of each storage location is compared with the risk field values ​​of its surrounding neighborhood locations to calculate the partial derivatives of the risk in the X, Y, and Z spatial dimensions, and then synthesizing the risk field change gradient vector. Another approach is to use machine learning or deep learning-based methods. For example, a convolutional neural network (CNN) model can be trained, inputting the real-time risk field value distribution map of all storage locations within the hazardous materials container. The convolutional kernels can then be used to extract features and detect edges in the risk field values, automatically identifying areas of drastic risk field value changes, i.e., the risk spread frontier.

[0152] Furthermore, based on the time-series variation data of the real-time field value within at least one risk diffusion frontier region, a risk evolution rate corresponding to each risk diffusion frontier region is fitted and generated. The risk evolution rate is an indicator that measures the speed and trend of risk change over time within a specific region. By fitting time-series data, it is possible to predict whether the risk will accelerate its spread, stabilize, or gradually subside in the future, providing a forward-looking basis for the dynamic adjustment of control policies. In practical applications, time series analysis methods can be used, such as Autoregressive Moving Average (ARMA), Long Short-Term Memory (LSTM), or Kalman filtering. The system collects real-time field value data of the risk diffusion frontier region over a past period, inputs this time-series data into a pre-trained time-series model, and the model learns historical patterns to fit the changing trend of the risk field value and predict its evolution rate at future moments. Alternatively, fitting methods based on physicochemical reaction kinetic models can also be used. Combining known chemical reaction rate equations and diffusion laws, the real-time field value within the risk diffusion frontier region is used as the initial condition. By solving partial differential equations or numerical simulations, the diffusion and intensity changes of the risk over a future period are predicted, thereby obtaining the risk evolution rate.

[0153] Furthermore, the real-time field value, the risk evolution rate, and multi-level security thresholds are input into the dynamic generation model of the access control policy. This model comprehensively calculates the current risk state and future risk trends, outputting and executing an access control policy bound to the current risk level. The dynamic generation model of the access control policy is an intelligent decision-making system that comprehensively considers the current risk level, future risk trends, and preset security standards to generate the most suitable access control policy. Its core lies in achieving intelligent linkage between risk perception and access control, ensuring that appropriate control measures can be taken under different risk scenarios. This model can be built based on a rule engine or expert system. A series of mapping rules between risk levels and access control policies are predefined, for example: "When the real-time field value exceeds the intermediate threshold and the risk evolution rate is on the rise, the access level is raised to the high level, and the corresponding authentication factor combination, reagent unlocking range, and monitoring sampling frequency are executed." After receiving the real-time field value, risk evolution rate, and multi-level security thresholds, the model matches these rules and outputs the corresponding access control policy. Another implementation method is to use a decision model based on reinforcement learning or deep learning. Through extensive training in a simulated environment, the model learns how to select the optimal access control strategy to minimize risk under different risk states and evolution trends. In actual operation, the model takes real-time field values ​​and risk evolution rates as input, performs inference through its internal neural network structure, and directly outputs an access control strategy that includes the combination of authentication factors, reagent unlocking range, and monitoring sampling frequency.

[0154] Through the above technical solution, this application overcomes the limitations of related technologies, such as inaccurate permission adjustments and the inability to respond promptly to rapidly escalating risk frontier regions or predict future risk escalation. Specifically, through spatial gradient analysis, the system can identify risk frontier regions where the risk field strength is rapidly increasing, enabling permission adjustments to be accurately targeted at high-risk areas, avoiding blind or delayed permission adjustments. Simultaneously, the introduction of risk evolution rate allows the system to predict future risk development trends, thereby taking stricter permission control measures before the risk fully erupts. These measures include increasing the complexity of authentication factor combinations, narrowing the reagent unlocking range, or increasing monitoring sampling frequency, effectively preventing dangerous operations and nipping potential risks in the bud. This dynamic permission adjustment mechanism based on spatial distribution and temporal evolution trends enhances the intelligence and proactive defense capabilities of hazardous chemical cabinet safety management, ensuring the safety of hazardous chemical retrieval processes.

[0155] In some of the above-mentioned schemes in this application, spatial gradient analysis is proposed to identify the risk diffusion frontier region. However, in this process, spatial gradient analysis may not fully consider the spatial distance and directional changes between storage locations, resulting in insufficient accuracy in the calculation of the risk field change gradient. This makes it impossible to accurately quantify the distribution differences and evolution direction of risks in the spatial dimension, thereby affecting the accuracy of identifying the risk diffusion frontier region.

[0156] To address this issue, this application proposes a method for spatial gradient analysis of the real-time field value of a four-dimensional spatiotemporal risk field, in order to obtain the risk field variation gradient of this real-time field value across different storage locations within a hazardous materials container. The method specifically includes the following steps: The risk field strength value of each storage location within the hazardous materials container at the current moment is extracted from the four-dimensional spatiotemporal risk field. This risk field strength value refers to a quantitative risk indicator formed by the combined effects of factors such as multi-reagent coupled risk matrices, spatial propagation, and time decay at a specific storage location within the hazardous materials container. Its magnitude directly reflects the potential hazard level currently faced at that location. Extracting this risk field strength value can be achieved by directly querying the four-dimensional spatiotemporal risk field data structure stored in a database or memory to obtain the risk value corresponding to the spatial coordinates and timestamp. Alternatively, it can be obtained by spatial interpolation or sampling of the four-dimensional spatiotemporal risk field to obtain the risk field strength value for a specific storage location.

[0157] Identify at least one neighboring storage location for each storage location. A neighboring storage location refers to other storage locations within the physical space of the hazardous materials cabinet that are adjacent to or within a certain distance of the current storage location, and where there is a potential risk propagation or mutual influence between these locations. One method for determining neighboring storage locations is based on a pre-defined cabinet structure topology map, defining physically adjacent (e.g., vertically, horizontally, vertically, frontally, and rearward) storage units as neighbors. Another method is to set a spatial distance threshold, identifying all other storage locations whose Euclidean or Manhattan distance from the current storage location is less than this threshold as neighbors.

[0158] Calculate the field strength difference between the risk field strength value at each storage location and the risk field strength values ​​at each corresponding neighboring storage location. This field strength difference refers to the numerical difference between the risk field strength value at the current storage location and the risk field strength value at any neighboring storage location. This difference quantifies the non-uniformity of risk in local space and is a direct reflection of risk diffusion or attenuation. This field strength difference can be calculated using simple arithmetic subtraction, i.e., subtracting the risk value at the neighboring location from the current location's risk value. Alternatively, a normalized difference can be used, such as dividing the difference by the average of the two risk values, to eliminate the influence of dimensions.

[0159] Based on this, and using the field strength difference and the spatial distance between each storage location and its corresponding neighboring storage locations, the directional gradient component of each storage location relative to its neighboring storage locations is calculated. This directional gradient component refers to the rate at which the risk field strength changes with spatial distance in a specific direction. It considers not only the difference in risk intensity but also incorporates the spatial distance factor, making the gradient calculation more consistent with the actual physical propagation laws. One way to calculate this directional gradient component is to divide the field strength difference by the spatial distance between the two storage locations and multiply it by the unit direction vector from the current location to the neighboring locations. Another way is to use the finite difference method, combined with the discretization characteristics of the spatial grid, to calculate the rate of change of risk along a specific axis and project it onto the direction connecting the two points.

[0160] Vector synthesis is performed on all directional gradient components corresponding to each storage location to obtain the risk field change gradient for each storage location. This vector synthesis refers to the comprehensive calculation of gradient components in all neighborhood directions surrounding a storage location to obtain the overall risk field change gradient vector for that location. The magnitude of this synthesized vector represents the severity of the risk change, and the direction represents the dominant direction of risk diffusion or attenuation. The synthesis method can be a simple vector summation, i.e., adding all directional gradient components together. Alternatively, a weighted average method can be used, assigning different weights based on the importance or distance of neighboring locations.

[0161] The risk field change gradient field corresponding to the real-time field value is generated by combining the risk field change gradients of all storage locations. This risk field change gradient field refers to the overall spatial distribution of the risk field change gradients of all storage locations within the hazardous materials container. It provides a global view of risk evolution trends, clearly showing the "flow" and "intensity" distribution of risks. This risk field change gradient field can be generated by storing the calculated risk field change gradient for each storage location in a multidimensional array or matrix corresponding to the spatial structure within the container. Alternatively, these gradient data can be associated with corresponding spatial coordinates to form a dataset available for visualization and further analysis.

[0162] Through the aforementioned technical solution, this application can accurately capture the spatial distribution differences and evolution direction of risks within hazardous chemical storage containers. Specifically, by extracting real-time risk field strength values ​​from a four-dimensional spatiotemporal risk field and combining them with the field strength differences and spatial distances of neighboring storage locations, directional gradient components are calculated. This ensures that the calculation of risk gradients is no longer a simple numerical difference but fully considers physical spatial characteristics, more accurately reflecting the propagation and attenuation effects of risks. Furthermore, by vector synthesis of all directional gradient components, the overall risk change trend of each storage location can be comprehensively and holistically characterized, avoiding bias in a single direction. Combining the risk field change gradients of all storage locations to generate a risk field change gradient field provides a high-precision and high-reliability input for subsequent identification of risk diffusion frontier regions, thereby improving the accuracy of risk perception and early warning, and providing a more solid data foundation for dynamically adjusting access control strategies.

[0163] In some of the above-mentioned schemes in this application, a risk evolution rate is generated by fitting real-time field values ​​to the temporal change data in the risk diffusion frontier region to predict future risk trends. However, in its implementation, there are deficiencies in how to accurately quantify the driving strength of newly added coupled risks on the risk field evolution based on the residual decay curve and concentration change trend, and how to combine the risk field change gradient and spatial range to accurately characterize the rate of risk spread to the neighborhood space, so as to generate a comprehensive and accurate risk evolution rate.

[0164] To address this, this application further proposes a method for fitting and generating a risk evolution rate corresponding to each risk diffusion frontier region based on the time-series variation data of real-time field values ​​within at least one risk diffusion frontier region. This method includes: obtaining the predicted coupled risk increment of the reagent combination corresponding to each risk diffusion frontier region at a future time from a preset risk prediction module; this predicted coupled risk increment is generated based on the remaining amount decay curves and concentration change trends of the at least two reagents to be used; determining the risk source term change rate for each risk diffusion frontier region based on the real-time field value and the predicted coupled risk increment; this risk source term change rate characterizes the driving strength of the newly added coupled risk on the risk field evolution; determining the risk diffusion propagation change rate for each risk diffusion frontier region based on the risk field change gradient within each risk diffusion frontier region and the spatial range pointed to by the risk field change gradient; this risk diffusion propagation change rate characterizes the rate at which the risk spreads to the neighboring space; and superimposing and fusing the risk source term change rate with the risk diffusion propagation change rate to obtain the risk evolution rate corresponding to each risk diffusion frontier region.

[0165] Specifically, the pre-defined risk prediction module can be a standalone software component or service. Its function is to predict potential future risks based on historical data, chemical reaction mechanisms, and reagent characteristics. For example, this module can employ machine learning-based models, such as recurrent neural networks (RNNs) or long short-term memory networks (LSTMs), to predict future risk trends by learning from historical risk evolution data and relevant environmental parameters. Alternatively, the module can infer potential risk increments based on expert systems and predefined chemical reaction rule bases, combined with the physicochemical properties of reagents. This predicted coupled risk increment refers to a quantitative indicator of the potential increase in the risk field value at a future point in time due to interactions between reagents. It is generated based on the decay curve and concentration change trend of the remaining amount of reagent to be used. For example, the change in the remaining amount can be predicted by modeling the consumption rate of the reagent and combining it with its volatilization, decomposition, or reaction rate under specific conditions. Simultaneously, the concentration distribution change of the reagent within the cabinet space can be predicted using diffusion models or reaction kinetic models. These curves and trends can be derived from experimental data, theoretical calculations, or fitting of historical monitoring data. The rate of change of the risk source term is used to quantify the driving strength of newly added coupled risks on the evolution of the risk field, that is, the rate at which the risk field value increases per unit time due to the emergence of new risk events or risk factors. It can be determined by comparing the current real-time field value with the predicted increment of coupled risks. For example, methods such as weighted averaging, differential calculation, or nonlinear mapping based on neural networks can be used to transform the predicted increment into a driving force on the current risk field. This rate of change reflects the inherent generation or enhancement trend of risk. The spatial range pointed to by the risk field change gradient refers to the physical area that may be affected by the risk in the direction of risk propagation indicated by the risk field change gradient. For example, based on the direction and magnitude of the risk field change gradient, combined with the layout of the storage units within the cabinet and the physical diffusion characteristics of the reagents (such as volatility and fluidity), the adjacent storage areas or specific spatial areas that the risk may spread to can be determined. Alternatively, a distance-based neighborhood range can be preset, and all areas within this range pointed to by the gradient are considered as the spatial range it points to. The rate of change of risk diffusion and propagation is used to characterize the rate at which risk spreads to the neighborhood space, that is, the speed and intensity of risk diffusion from high-risk areas to low-risk areas. The determination of this rate of change can be achieved by calculating the flux of the risk field change gradient within a specific spatial range. For example, the diffusion rate of the risk value in space can be calculated using methods based on the finite difference or finite element method. Alternatively, the propagation rate of the risk per unit time through that range can be estimated based on the magnitude of the risk field change gradient and the physical characteristics of the spatial range it points to (such as spatial capacity and obstacles). This rate of change focuses on the external spread of the risk in space. The superposition and fusion refers to combining the rate of change of the risk source term with the rate of change of risk diffusion and propagation to obtain a comprehensive risk evolution rate. For example, a simple linear superposition can be used, i.e., directly adding the two together.Alternatively, a weighted superposition method can be used, assigning different weights to the source term-driven and diffusion-propagation factors based on their relative importance to risk evolution in the actual scenario. Or, a nonlinear fusion function, such as a model based on fuzzy logic or a neural network, can be used to more accurately reflect the complex interaction between the two.

[0166] Through the above technical solution, this application can accurately predict the future evolution trend of risks within hazardous chemical storage containers. Specifically, by obtaining the predicted coupled risk increment generated based on the remaining amount decay curve and concentration change trend from the preset risk prediction module, the system can fully consider the impact of actual reagent consumption and dynamic changes in environmental concentration on future risks, thereby improving the reliability and pertinence of risk increment prediction. Based on this, by combining real-time field values ​​and predicted coupled risk increments to determine the risk source term change rate, the system can accurately capture the driving strength of newly added coupled risks on the risk field evolution, making the perception of changes in risk sources more sensitive. Simultaneously, based on the risk field change gradient and its spatial range, the risk diffusion and propagation change rate is determined, quantifying the rate of risk spread to the adjacent space and considering the actual path and physical characteristics of risk propagation. By superimposing and fusing the risk source term change rate with the risk diffusion and propagation change rate, a comprehensive and accurate risk evolution rate is generated. This rate not only reflects the generation and enhancement of risks but also considers their spatial diffusion, providing a more accurate and forward-looking basis for subsequent adjustments to access control strategies. This enables the system to shift from passive response to proactive prediction and intervention, improving the intelligence and safety of hazardous chemical container risk management.

[0167] In some of the above-mentioned schemes in this application, the rate of change of risk source terms is determined based on real-time field values ​​and predicted coupled risk increments to calculate the risk evolution rate. However, in this process, there is a lack of accuracy in separating the current risk contribution and quantifying the dynamic impact of future risk increments on the changes of risk source terms, which limits the accuracy of risk evolution prediction.

[0168] To address this, this application further proposes a method for determining the rate of change of risk source terms in each risk diffusion frontier region based on the real-time field value and the predicted coupling risk increment within each region. This method includes: performing source term decomposition on the real-time field value within each risk diffusion frontier region, extracting the current source term component contributed by reagent coupling occurring at the current moment; performing time-series expansion on the predicted coupling risk increment to obtain an increment sequence at multiple future moments, and fitting an increment contribution decay curve based on this increment sequence; determining the source term superposition value for each risk diffusion frontier region per unit time in the future based on the current source term component and the increment contribution decay curve; and subtracting the current source term component from the superposition value to obtain the rate of change of risk source terms in each risk diffusion frontier region.

[0169] Specifically, source term decomposition is performed on the real-time field values ​​within each risk diffusion frontier region. The aim is to precisely separate the risk component directly contributed by the currently occurring reagent coupling reaction from the total observed risk intensity, thereby avoiding interference from historical or external factors and ensuring the timeliness and accuracy of risk source term calculations. Source term decomposition can be achieved in ways including, but not limited to: one approach is based on physicochemical models, utilizing known chemical reaction kinetic equations and thermodynamic data, combined with real-time monitored parameters such as reagent type, concentration, and temperature, to separate the portion directly contributed by the current reagent coupling reaction from the real-time field value through inverse calculation or model fitting. Another approach is based on machine learning models, training a machine learning model (such as a neural network or support vector machine), inputting real-time field values, historical reagent coupling data, and environmental parameters, and outputting the source term component contributed by reagent coupling at the current moment. This model, by learning patterns in historical data, can identify and separate the contributions of different risk sources.

[0170] The predicted coupling risk increment is subjected to time-series expansion to obtain an increment sequence at multiple future time points. An increment contribution decay curve is then fitted based on this increment sequence. The predicted coupling risk increment is an estimate of potential future coupling risks based on the decay curve of the remaining amount of reagent to be used and its concentration change trend. Time-series expansion refines this overall future risk increment into a series of discrete increment values ​​corresponding to different future time points, thereby capturing the dynamic process of risk evolution. The increment contribution decay curve is used to quantify the natural decay trend of the contribution of these future increments to the total risk over time. Methods for implementing time-series expansion and curve fitting include, but are not limited to: one method is based on an exponential decay model, assuming that the contribution of the predicted coupling risk increment decays exponentially over time, and determining the optimal decay parameters through nonlinear regression analysis of the increment sequence. Another method is based on a polynomial fitting model; when the decay trend of the risk increment is more complex, a polynomial function can be used to fit the increment sequence to capture more complex decay patterns.

[0171] Based on this, and using the current source term component and the incremental contribution decay curve, the source term superposition value for each risk diffusion frontier region in the future unit time is determined. This step aims to integrate the current risk contribution with the expected future risk increment to form a forward-looking assessment of the total risk contribution in the future unit time. Methods for determining the source term superposition value can include, but are not limited to: one method is to use integration or summation calculations, superimposing the results of the integration (for continuous models) or summation (for discrete models) of the current source term component and the incremental contribution decay curve in the future unit time. Another method is to use a predictive model, taking the parameters of the current source term component and the incremental contribution decay curve as input, and directly outputting the source term superposition value in the future unit time.

[0172] Subtracting the current source term component from the summed value of the source terms yields the rate of change of risk source terms in each risk diffusion frontier region. This difference calculation directly quantifies the net change of risk source terms within a unit of time in the future, i.e., the instantaneous trend of risk source terms. Methods for calculating the rate of change of risk source terms can include, but are not limited to: one method is to directly perform the difference calculation, i.e., calculate according to the formula "Rate of change of risk source terms = Summed value of source terms - Current source term component". Another method is to perform normalization processing based on the direct difference calculation to facilitate comparison between different risk scenarios or as input to other models.

[0173] Through the above technical solution, this application can accurately calculate the rate of change of risk source terms, effectively solving the problem of insufficient dynamic quantification in risk prediction. Source term decomposition of real-time field values ​​isolates reagent coupling contributions occurring at the current moment, avoiding interference from historical data and ensuring the timeliness and accuracy of risk source term calculation. Temporal unfolding of the predicted coupled risk increment decomposes future risks into multiple increment sequences, capturing the dynamics of risk changes and providing structured data for subsequent analysis. An increment contribution decay curve is generated by fitting the increment sequence, quantifying the decay trend of risk increments, reflecting the natural law of risk evolution over time, and enhancing the reliability of prediction. Then, based on the current source term components and the increment contribution decay curve, the superimposed value of source terms per unit time in the future is determined, integrating current and future risk contributions to achieve a forward-looking assessment of the total risk. Subtracting the current source term component from the superimposed value directly calculates the change, accurately characterizing the instantaneous rate of change of risk source terms and providing key input for risk evolution. The entire process, through decomposition, temporal analysis, and difference calculation, improves the quantitative accuracy of the rate of change of risk source terms, thereby optimizing the risk prediction effect. This precise rate of change of risk sources, as a key component of the risk evolution rate, enables more dynamic and accurate access control strategies for hazardous chemical storage facilities. This allows for more timely and effective adjustment of permissions when risks escalate, preventing potentially dangerous operations and improving the safety of hazardous chemical management.

[0174] In some of the schemes mentioned above in this application, the risk evolution rate is determined to quantify the risk spread trend. However, in the process of implementation, there are shortcomings in how to accurately calculate the rate of change of risk diffusion and propagation, especially considering spatial capacity and risk flux.

[0175] To address this, this application further proposes a method for determining the risk diffusion propagation rate of each risk diffusion frontier region based on the risk field change gradient and the spatial range to which the risk field change gradient points. This process includes: determining at least one propagation target region to which the risk field change gradient points based on the risk field change gradient within each risk diffusion frontier region; obtaining a spatial capacity parameter for at least one propagation target region from the cabinet's spatial location data, where the spatial capacity parameter characterizes the physical space size and reagent density that each propagation target region can accommodate; determining the risk flux input from the risk field change gradient to each propagation target region based on the magnitude of the risk field change gradient and the spatial capacity parameter; and summing the risk flux input from each risk diffusion frontier region to all propagation target regions to obtain the risk diffusion propagation rate of each risk diffusion frontier region.

[0176] Specifically, the risk diffusion frontier region refers to the area within the hazardous materials container where the risk field strength is rapidly increasing or about to increase. It represents the boundary or hotspot of risk spread, and its function is to identify the starting point or key area of ​​risk propagation for centralized monitoring and assessment. For example, spatial gradient analysis of the risk field strength can identify areas where the gradient value exceeds a preset threshold. Alternatively, combining historical risk evolution data and machine learning models can predict areas where future risks may grow rapidly. The risk field change gradient characterizes the rate and direction of spatial change of the risk field strength within the hazardous materials container, indicating the direction and intensity of risk propagation and serving as an important indicator of risk diffusion trends. For example, the difference in risk field strength between adjacent storage locations can be calculated using the finite difference method or convolution kernel operations, and a vector can be synthesized. Alternatively, deep learning models can be used to learn and extract the spatial variation characteristics of the risk field strength from high-dimensional risk data. The propagation target region refers to the adjacent storage area that may be affected by the risk, pointed to by the risk field change gradient of the risk diffusion frontier region. Its function is to clarify the specific spatial range to which the risk may spread, providing a target for subsequent risk flux calculations. For example, the directly adjacent storage unit pointed to by the direction vector of the risk field change gradient can be determined. Alternatively, by combining the cabinet layout diagram and a pre-defined diffusion path model, a series of potentially affected areas along the gradient direction can be identified. This cabinet spatial location data describes the geometric coordinates, physical dimensions, and adjacency relationships of each storage location within the hazardous materials cabinet. Its function is to provide information on the physical carriers of risk propagation within the cabinet space, forming the basis of spatial analysis. For example, a pre-established 3D model or 2D planar layout diagram of the cabinet can be used to store the ID, coordinates, length, width, height, and a list of adjacent storage compartments for each compartment. Alternatively, a sensor network can be used to acquire the location information of reagent bottles in real time, and this can be combined with cabinet structural data for spatial mapping. This spatial capacity parameter quantifies the physical space size and reagent density that each propagation target area can accommodate. Its function is to reflect the target area's ability to "buffer" or "amplify" risk, affecting the calculation of risk flux. For example, the physical space size can be directly calculated from the volume or area of ​​the storage compartment. The reagent density can be obtained by statistically analyzing the ratio of the number or total mass of reagent bottles in the current compartment to the compartment volume. Alternatively, by combining the physicochemical properties of the reagents and the airtightness of the storage container, a weighted correction can be made to the physical space size and reagent density to obtain more refined spatial capacity parameters. The magnitude of the risk field change gradient is the size of the risk field change gradient vector, representing the drastic change in risk field strength in space. Its role is to quantify the intensity of risk diffusion and is a key factor in calculating risk flux. For example, it can be measured by calculating the Euclidean norm of the risk field change gradient vector. Alternatively, other norms such as Manhattan distance or Chebyshev distance can be used to measure the gradient strength.Risk flux refers to the amount of risk passing through a cross-section or entering a region per unit time. Specifically, it refers to the risk input from the risk field gradient to the target region, quantifying the actual intensity of risk propagation from the source region to the target region. For example, the magnitude of the risk field gradient can be multiplied by the spatial capacity parameter, incorporating a time factor. Alternatively, a risk flux calculation formula based on fluid dynamics or heat conduction models can be constructed, considering factors such as diffusion coefficients and concentration differences. Accumulation refers to summing multiple values, comprehensively considering the risk input from a risk diffusion front to all possible target regions, yielding the overall risk diffusion and propagation change rate. For example, simple arithmetic summation can be used. Alternatively, a weighted summation can be performed based on the importance or distance of different target regions. The resulting risk diffusion and propagation change rate characterizes the speed and intensity of risk spreading to its neighborhood in the spatial dimension, and is an important component of the risk evolution rate, used to predict the future spatial distribution of risk.

[0177] Through the above technical solutions, the rate of change of risk diffusion and propagation can be quantified more accurately when determining the risk evolution rate. Specifically, by determining the propagation target area based on the risk field change gradient, the identification of risk propagation paths becomes more accurate, avoiding invalid calculations for irrelevant areas, thereby improving computational efficiency and accuracy. Furthermore, by obtaining spatial capacity parameters from the spatial location data within the cabinet and using them to characterize the physical space size and storage reagent density that each propagation target area can accommodate, this application fully considers the actual constraints and impacts of the internal environment of the hazardous chemical cabinet on risk diffusion, solving the risk assessment bias problem caused by the neglect of spatial factors in traditional methods. On this basis, the risk flux is determined jointly based on the modulus of the risk field change gradient and the spatial capacity parameter, which can accurately quantify the actual input intensity of risk from high-risk areas to potentially affected areas, making the intensity of risk propagation reflect the real situation under spatial constraints. By accumulating the risk flux input from a risk diffusion front area to all propagation target areas, multi-directional risk spread is comprehensively considered, avoiding the limitations of single-directional assessment, thus providing a more comprehensive and accurate spatial diffusion component for subsequent calculation of the risk evolution rate. Combined with the risk source change rate proposed in the above scheme, the risk diffusion and propagation change rate provided by this application can more comprehensively and accurately depict the evolution trend of risk in space and time, providing a solid data foundation for the fitting generation of risk evolution rate. This enables the dynamic generation model of permission policy to output a more reasonable and dynamic permission control policy based on more accurate risk prediction, effectively improving the safety management level of hazardous chemical cabinets.

[0178] In some of the solutions mentioned above in this application, a dynamic generation model of permission strategy is proposed to comprehensively calculate the current risk status and future risk trends to output permission control strategies. However, in this process, there are shortcomings in how to dynamically adjust the combination of authentication factors, reagent unlocking range and monitoring sampling frequency based on the changing trend of risk level to ensure that the permission strategy is synchronized with the risk evolution. This may lead to the strategy adjustment lagging behind the risk development or lack of predictive response capability.

[0179] To address this, this application further proposes a dynamic generation model using permission policies to comprehensively calculate the current risk status and future risk trends, outputting and executing permission control policies bound to the current risk level. This method fuses real-time field values ​​and risk evolution rates to obtain the predicted risk field value for each risk diffusion frontier region at a predetermined future time. The real-time field value reflects the current risk distribution within the hazardous materials container, while the risk evolution rate characterizes the spatial and temporal trends of risk change. By fusing these two factors, the system can proactively predict the risk status at a future point in time, combining the current risk distribution and its changing trends, thereby enabling early perception and response to potential risks. For example, a weighted average model can be used, with the real-time field value as the benchmark and the risk evolution rate as the correction term, linearly or non-linearly superimposed using predetermined weighting factors to obtain the predicted risk field value. Alternatively, a machine learning-based time series prediction model, such as the Autoregressive Integral Moving Average (ARIMA) model or Long Short-Term Memory (LSTM) network, can be used, with historical real-time field values ​​and risk evolution rates as input to train the model and predict future risk field values.

[0180] The real-time risk field value and the predicted risk field value are compared with the multi-level safety threshold, and the current risk level and the predicted risk level are determined based on the comparison results. This step aims to map the quantified risk field value (including real-time and predicted) to discrete risk levels so that the system can make decisions and adjust strategies. Specifically, multiple discrete threshold intervals can be set. For example, the risk field value in the range [0, T1) can be defined as low risk, [T1, T2) as medium risk, [T2, T3) as high risk, and [T3, +∞) as extremely high risk. The corresponding risk level is determined by judging which interval the risk field value falls into. Alternatively, a fuzzy logic reasoning system can be used, taking the risk field value as fuzzy input, and outputting the corresponding fuzzy risk level through predefined fuzzy rules and membership functions, and then performing defuzzification processing to obtain the accurate risk level.

[0181] Based on this, and considering the risk level transition relationship between the current and predicted risk levels, the system retrieves the corresponding strategy generation rule from the dynamic generation model of the permission policy to obtain a combination of permission control parameters, including authentication factor combinations, reagent unlocking ranges, and monitoring sampling frequencies. The risk level transition relationship describes the trend (e.g., increasing, decreasing, or remaining unchanged) and magnitude of risk level changes from the present to the future. According to this relationship, the system can intelligently select the most suitable strategy generation rule to ensure the accuracy and adaptability of permission adjustments. For example, a strategy rule library can be pre-set, containing strategy templates or adjustment rules for different risk level transition relationships (e.g., "low risk -> medium risk", "medium risk -> high risk", "high risk -> low risk", etc.). The system directly retrieves and calls the corresponding rule from the library based on the identified risk level transition relationship. Alternatively, an expert system based on decision trees or rule engines can be used, taking the current and predicted risk levels as input, and dynamically generating or selecting strategy generation rules through a series of pre-set decision logics and conditional judgments.

[0182] The access control parameter combination is sent to the execution module, which then authenticates the current user according to the authentication factor combination, unlocks the electronic lock array of the hazardous materials cabinet according to the reagent unlocking range, and dynamically adjusts the sampling frequency of the sensor network according to the monitoring sampling frequency. This is a crucial step in translating decision-making into action, ensuring that the system can adjust user permissions, equipment status, and monitoring behavior in real time according to changes in risk level, thereby effectively controlling risk. For example, regarding the authentication factor combination, fingerprint recognition may be sufficient at low risk, a combination of fingerprint recognition and password may be required at medium risk, and fingerprint recognition, facial recognition, and remote authorization from management personnel may be required at high risk. Regarding the reagent unlocking range, all requested reagents can be unlocked at low risk, only some low-risk reagents may be unlocked at medium risk, and unlocking may be prohibited or require on-site authorization from management personnel at high risk. Regarding the monitoring sampling frequency, it can be set to sample once per minute at low risk, adjusted to once every 30 seconds at medium risk, and further increased to once every 5 seconds at high risk to obtain more intensive real-time data. The execution module can be the central controller of the hazardous materials cabinet, which communicates with the identity authentication system, electronic lock array and sensor network to issue instructions and receive feedback.

[0183] Through the above technical solution, this application can deeply integrate real-time risk status with future risk trends, realizing dynamic and forward-looking adjustments to the access control strategy for hazardous chemical storage cabinets. Specifically, by integrating real-time risk field values ​​and risk evolution rates, the system can predict future risk field values, thus anticipating their development trends before they fully manifest. Comparing real-time and predicted risk field values ​​with multi-level safety thresholds quantifies current and future risk levels, providing a clear basis for adjusting access control strategies. Based on the transition relationship between current and predicted risk levels, the system can intelligently invoke matching strategy generation rules to ensure that the generated combinations of authentication factors, reagent unlocking ranges, and monitoring sampling frequencies accurately respond to rising or falling risk trends, avoiding lag in strategy adjustments. By distributing these dynamically adjusted access control parameter combinations to the execution module, the system can adjust user authentication, the electronic lock array of the hazardous chemical storage cabinet, and the sampling frequency of the sensor network in real time. This allows for timely tightening of access control when risks increase to prevent potentially dangerous operations, and appropriate relaxation of access control when risks decrease to improve operational convenience. This dynamic permission linkage mechanism based on risk evolution trends improves the level of intelligent safety management of hazardous chemical cabinets and the timeliness and effectiveness of risk response. It effectively solves the problems of disconnect between permission policies and real-time risk status, delayed adjustments, or lack of predictive response capabilities in traditional solutions.

[0184] In some of the embodiments described above in this application, an access control strategy is proposed to dynamically adjust permissions based on the current risk status and future risk trends. However, in its implementation, the strategy adjustment may not be refined enough and may not be able to adaptively adjust according to the trend and magnitude of changes in risk level, resulting in inaccurate access control or untimely response.

[0185] In response, this application further proposes, based on the aforementioned level transition relationship between the current risk level and the predicted risk level, to call the strategy generation rule corresponding to the level transition relationship from the dynamic generation model of the permission policy, thereby obtaining a combination of permission control parameters including the authentication factor combination method, reagent unlocking range, and monitoring sampling frequency, including: This analysis examines the direction and magnitude of the risk transition between the current and predicted risk levels. The direction of the transition indicates whether the risk is trending upward or downward, while the magnitude of the transition indicates the severity of the risk change.

[0186] Based on the direction and magnitude of the level transition, a policy adjustment template matching the direction and magnitude of the level transition is queried from the dynamic generation model of the permission policy. The policy adjustment template includes adjustment rules for the combination of authentication factors, adjustment rules for the reagent unlocking range, and adjustment rules for the monitoring sampling frequency.

[0187] Get the current combination of permission control parameters that are currently being executed.

[0188] The current combination of access control parameters is incrementally adjusted according to the strategy adjustment template to obtain the combination of access control parameters corresponding to the level transition relationship.

[0189] Specifically, when analyzing the direction and magnitude of the risk level transition between the current and predicted risk levels, the predicted risk level value can be compared with the current risk level value. For example, if the predicted risk level value is higher than the current risk level value, the direction of the transition is determined to be an upward trend. Conversely, if the predicted risk level value is lower than the current risk level value, the direction of the transition is determined to be a downward trend. Meanwhile, the magnitude of the transition can be quantified by calculating the difference or ratio between the predicted and current risk levels. For example, multiple magnitude ranges (such as weak, moderate, and severe) can be set, and the calculated difference or ratio can be mapped to the corresponding ranges to characterize the severity of the risk change. Another approach is to predict the future trend and quantify the magnitude of the change by analyzing the rate of change of the risk level within a preset time window, combined with statistical methods (such as moving averages and exponential smoothing).

[0190] When querying a dynamic generation model of permission policies based on the direction and magnitude of a rank transition, the model can pre-define a multi-dimensional lookup table or decision tree. The input dimensions of this lookup table include the rank transition direction and magnitude, and the output is the corresponding policy adjustment template identifier or the template content itself. For example, when the risk shows a "moderate upward" trend, the system will match the pre-defined "moderate risk increase adjustment template." When the risk shows a "sharp downward" trend, it will match the "sharp risk decrease adjustment template." Alternatively, the dynamic generation model of permission policies can also employ a rule-based expert system to match and select policy adjustment templates according to predefined logical rules (e.g., IF rank transition direction is upward AND rank transition magnitude is sharp THEN select template A).

[0191] When retrieving the currently executing combination of access control parameters, the system can maintain a real-time updated state variable for this parameter combination. This variable is updated after each access control policy execution and stores parameters such as the currently effective authentication factor combination, reagent unlocking range, and monitoring sampling frequency. This allows for quick retrieval and serves as the basis for policy adjustments when needed. Alternatively, the access control module can persist its currently effective parameter combinations to a database or configuration center with a timestamp each time an access control policy is executed, enabling subsequent queries and retrieval of the latest or specific time-point parameter combinations.

[0192] When incrementally adjusting the current access control parameter combination according to the policy adjustment template, the policy adjustment template can contain specific adjustment instructions for each parameter, such as the authentication factor combination method, reagent unlocking range, and monitoring sampling frequency. For example, the template might indicate "Authentication factor combination method: Add fingerprint authentication to the existing method," "Reagent unlocking range: Narrow to allow only single-bottle reagent access," and "Monitoring sampling frequency: Increase to once per second." The system parses these instructions and applies them to the current access control parameter combination to achieve incremental parameter modification. Furthermore, the policy adjustment template can also define adjustment functions or incremental values ​​for parameters. For example, for the monitoring sampling frequency, the template might indicate "Increase sampling frequency by 20%" or "Adjust sampling frequency to high-frequency mode." The system calculates the new parameter values ​​based on the current values ​​and adjustment rules.

[0193] Through the above technical solution, this application can refine and adaptively adjust the access control strategy based on the direction and magnitude of the risk level transition between the current and predicted risk levels. By analyzing the trend and severity of risk level changes, precise input is provided for subsequent strategy adjustments, ensuring that the access control strategy can adapt to risk increases or decreases and the specific degree of change, avoiding blind responses. Based on the analyzed risk level transition direction and magnitude, a matching strategy adjustment template is queried, and predefined adjustment rules in the template are used to ensure that the adjustment process is tailored to the specific characteristics of risk changes, avoiding adjustment deviations caused by rule mismatches. Simultaneously, by obtaining the current combination of access control parameters being executed at the current moment and making incremental adjustments based on this, the strategy adjustment is made continuous, preventing instability caused by parameter mutations. This dynamic and refined access control method improves the system's timeliness and accuracy in responding to the risks of hazardous chemical container operations, thereby enhancing the overall adaptability and security of access control and effectively solving the problem of insufficiently refined strategy adjustments.

[0194] All of the above-mentioned optional technical solutions can be combined in any way to form the optional embodiments of this application, and will not be described in detail here.

[0195] Figure 6 This is a schematic diagram of a multi-reagent bulk retrieval and access control system for hazardous chemical cabinets provided in an embodiment of this application. See also... Figure 6 The system includes: The retrieval module 601 is used to respond to a batch retrieval request containing at least two reagents to be retrieved, retrieve the chemical kinetic parameters corresponding to the reagents to be retrieved from a pre-stored chemical characteristic parameter library, the chemical kinetic parameters including the reaction activation energy and the reaction heat, and apply the temperature parameter from real-time environmental monitoring data to the reaction activation energy to obtain the temperature-corrected activation energy parameter.

[0196] The calculation module 602 is used to calculate the coupling risk coefficient of at least two reagents to be taken in real time based on the temperature-corrected activation energy parameter, the heat of reaction and the real-time environmental monitoring data, and to aggregate the calculation results of all reagent combinations to generate a multi-reagent coupling risk matrix. The multi-reagent coupling risk matrix is ​​used to quantitatively characterize the chemical incompatibility risk of all reagent combinations in the batch take-up list in the real-time environment.

[0197] The input module 603 is used to input the multi-reagent coupled risk matrix and the cabinet spatial location data obtained from the sensor network into the spatiotemporal risk field evolution model. The spatiotemporal risk field evolution model is used to iteratively solve the distribution and evolution trend of risk in the spatial and temporal dimensions, and output the four-dimensional spatiotemporal risk field under the current operation scenario.

[0198] The execution module 604 is used to obtain and execute the access control strategy bound to the current risk level based on the comparison results of the real-time field value of the four-dimensional spatiotemporal risk field and the multi-level security threshold. The access control strategy includes the linkage adjustment of the authentication factor combination method, reagent unlocking range and monitoring sampling frequency.

[0199] It should be noted that the above-described embodiment of the hazardous chemical cabinet multi-reagent batch retrieval and access control system is only illustrated by the division of the functional modules described above. In practical applications, the functions can be assigned to different functional modules as needed, that is, the internal structure of the computer equipment can be divided into different functional modules to complete all or part of the functions described above. Furthermore, the hazardous chemical cabinet multi-reagent batch retrieval and access control system and the hazardous chemical cabinet multi-reagent batch retrieval and access control method embodiment belong to the same concept, and their specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0200] In an exemplary embodiment, a computer-readable storage medium is also provided, such as a memory including a computer program, which can be executed by a processor to complete the method for batch retrieval of multiple reagents and access control of hazardous chemical cabinets in the above embodiments. For example, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), magnetic tape, floppy disk, and optical data storage device, etc.

[0201] In an exemplary embodiment, a computer program product or computer program is also provided, which includes program code stored in a computer-readable storage medium. The processor of a computer device reads the program code from the computer-readable storage medium and executes the program code, causing the computer device to execute the above-described method for batch retrieval of multiple reagents and authorization linkage for hazardous chemical cabinets.

[0202] In some embodiments, the computer program involved in the present application embodiments may be deployed and executed on a computer device, or executed on multiple computer devices located in one location, or executed on multiple computer devices distributed in multiple locations and interconnected through a communication network. Multiple computer devices distributed in multiple locations and interconnected through a communication network may constitute a blockchain system.

[0203] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0204] The above are merely optional embodiments of this application and are not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for batch dispensing of multiple reagents from a hazardous chemicals cabinet and linking access permissions, characterized in that, The method includes: In response to a bulk retrieval request containing at least two reagents to be retrieved, chemical kinetic parameters corresponding to the reagents to be retrieved are retrieved from a pre-stored chemical characteristic parameter library. The chemical kinetic parameters include reaction activation energy and reaction heat. The reaction activation energy is corrected by applying a temperature parameter from real-time environmental monitoring data to obtain a temperature-corrected activation energy parameter. Based on the temperature-corrected activation energy parameter, the heat of reaction, and the real-time environmental monitoring data, the coupling risk coefficients of the at least two reagents to be taken are calculated in real time for pairwise combinations, and the calculation results of all reagent combinations are aggregated to generate a multi-reagent coupling risk matrix. The multi-reagent coupling risk matrix is ​​used to quantitatively characterize the chemical incompatibility risk of all reagent combinations in the batch take-up list under real-time environment. The multi-reagent coupled risk matrix and the cabinet spatial location data obtained from the sensor network are input into the spatiotemporal risk field evolution model. The spatiotemporal risk field evolution model is used to iteratively solve the distribution and evolution trend of risk in the spatial and temporal dimensions, and output the four-dimensional spatiotemporal risk field under the current operation scenario. Based on the comparison results between the real-time field value of the four-dimensional spatiotemporal risk field and the multi-level security threshold, an access control strategy bound to the current risk level is obtained and executed. The access control strategy includes the linkage adjustment of the authentication factor combination method, reagent unlocking range and monitoring sampling frequency.

2. The method according to claim 1, characterized in that, In response to a bulk retrieval request containing at least two reagents to be retrieved, the chemical kinetic parameters corresponding to the reagents to be retrieved are retrieved from a pre-stored chemical characteristic parameter library, including: In response to a bulk retrieval request containing at least two reagents to be retrieved, the reagent identifiers of the at least two reagents to be retrieved and the reagent category corresponding to each reagent are identified based on the bulk retrieval request; Based on the reagent identifier and the reagent category, the basic reaction activation energy and basic reaction heat corresponding to each reagent to be used are queried from the chemical characteristic parameter library; When the activation energy data of a reaction that is missing between any two reagents to be used is found to be missing, the activation energy is estimated by calling the pre-stored group contribution method estimation rule in the chemical feature parameter library based on the reagent category of the two reagents to be used, and combining the molecular structure characteristics of the two reagents to be used. The basic reaction activation energy, the basic reaction heat, and the estimated activation energy are integrated to generate a set of chemical kinetic parameters corresponding to the batch retrieval request.

3. The method according to claim 1, characterized in that, Based on the temperature-corrected activation energy parameter, the heat of reaction, and the real-time environmental monitoring data, the coupling risk coefficients of the at least two reagents to be used are calculated in real time for pairwise combinations. The calculation results for all reagent combinations are then aggregated to generate a multi-reagent coupling risk matrix, including: The temperature-corrected activation energy parameter is subjected to an exponential domain transformation to obtain the reaction potential factor corresponding to each reagent pair. The reaction potential factor is used to characterize the thermodynamic tendency of the chemical reaction to occur under the current temperature conditions. The reaction potential energy factor and the corresponding reaction heat are nonlinearly coupled to obtain the reaction intensity base value for each reagent combination. The reaction intensity base value is used to quantify the comprehensive intensity of the chemical reaction in both thermodynamic driving and energy release dimensions. Extract real-time concentration data and remaining amount data corresponding to the at least two reagents to be used from the real-time environmental monitoring data, and map the real-time concentration data and the remaining amount data together as a dynamic weight of reaction probability. The dynamic weight of reaction probability is used to characterize the synergistic contribution of the actual reactant mass and the reaction contact probability to the degree of risk. Based on the reaction intensity baseline, the reaction probability dynamic weight, and the cross-coupling compensation mechanism among multiple reagents, the coupling risk coefficient is comprehensively calculated and matrix-aggregated to generate the multi-reagent coupling risk matrix. The cross-coupling compensation mechanism is used to quantify the catalytic or inhibitory effect of the third type of reagent among the at least two reagents to be used on the current reagent combination.

4. The method according to claim 3, characterized in that, The process of comprehensively calculating and matrix-aggregating the coupling risk coefficient based on the reaction intensity baseline, the dynamic weight of the reaction probability, and the cross-coupling compensation mechanism among multiple reagents to generate the multi-reagent coupling risk matrix includes: Based on the list of at least two reagents to be taken, an initial coupling risk matrix is ​​constructed, wherein each element of the initial coupling risk matrix corresponds to the initial position of a pair of reagent combinations; The reaction intensity base value corresponding to each reagent pair is multiplied by the reaction probability dynamic weight to obtain the basic coupling risk value of each reagent pair, and the basic coupling risk value is filled into the corresponding matrix element position of the initial coupling risk matrix. The cross-coupling compensation matrix corresponding to the at least two reagents to be used is retrieved from the chemical feature parameter library. Each compensation element of the cross-coupling compensation matrix is ​​used to quantify the catalytic or inhibitory effect of the third type of reagent on the corresponding reagent combination. The initial coupling risk matrix after filling is multiplied element-wise with the cross-coupling compensation matrix, and the basic coupling risk value of each reagent combination is amplified by catalytic effect or attenuated by inhibitory effect to obtain the multi-reagent coupling risk matrix.

5. The method according to claim 1, characterized in that, The process involves inputting the multi-reagent coupled risk matrix and the cabinet spatial location data obtained from the sensor network into a spatiotemporal risk field evolution model. This model iteratively solves for the distribution and evolution trend of risks in both spatial and temporal dimensions, outputting a four-dimensional spatiotemporal risk field under the current operational scenario, including: The multi-reagent coupling risk matrix is ​​used as a risk source term and input together with the spatial location data inside the cabinet into the spatiotemporal risk field evolution model to initialize the initial risk distribution field of the spatiotemporal risk field evolution model at the current moment. The initial risk distribution field is subjected to neighborhood diffusion calculation by the spatial propagation operator in the spatiotemporal risk field evolution model to obtain the spatial propagation components of the risk between different spatial locations within the cabinet. The initial risk distribution field is subjected to time-series evolution calculation by the time decay operator in the spatiotemporal risk field evolution model to obtain the natural decay component of risk over time. Based on the spatial propagation component, the time decay component, and the newly added reagent coupling risk matrix input at the next moment, iterative fusion calculation is performed to output the four-dimensional spatiotemporal risk field under the current operation scenario.

6. The method according to claim 5, characterized in that, The step of performing neighborhood diffusion calculations on the initial risk distribution field using the spatial propagation operator in the spatiotemporal risk field evolution model to obtain the spatial propagation components of the risk between different spatial locations within the cabinet includes: Based on the spatial location data inside the cabinet, a set of neighboring locations for each storage location inside the hazardous chemical cabinet is determined. The set of neighboring locations includes other storage locations that are physically adjacent to the current storage location. The set of diffusion coefficients corresponding to the spatial propagation operator is retrieved from the spatiotemporal risk field evolution model. The set of diffusion coefficients includes the risk diffusion weight between the current storage location and each neighboring location in the set of neighboring locations. Extract the current risk value of the current storage location and the neighborhood risk value of each neighborhood location in the neighborhood location set from the initial risk distribution field; The neighborhood risk value and the corresponding risk diffusion weight are weighted and summed, and the weighted summation result is fused with the current risk value to obtain the spatial propagation component of the current storage location.

7. The method according to claim 1, characterized in that, The comparison result between the real-time field value of the four-dimensional spatiotemporal risk field and the multi-level security thresholds yields and executes an access control policy bound to the current risk level, including: Spatial gradient analysis is performed on the real-time field value of the four-dimensional spatiotemporal risk field to obtain the risk field change gradient between different storage locations within the cabinet, and at least one risk diffusion front region with rapidly increasing risk field strength is identified based on the risk field change gradient. Based on the time-series change data of the real-time field value in at least one risk diffusion frontier region, a risk evolution rate corresponding to each risk diffusion frontier region is fitted and generated. The risk evolution rate is used to characterize the spread speed and intensity growth trend of the risk in the future. The real-time field value, the risk evolution rate, and the multi-level security threshold are input into the permission policy dynamic generation model. The permission policy dynamic generation model is used to comprehensively calculate the current risk status and future risk trends, output the permission control policy bound to the current risk level, and execute it.

8. The method according to claim 7, characterized in that, The step of fitting and generating a risk evolution rate corresponding to each risk diffusion frontier region based on the time-series change data of the real-time field value within the at least one risk diffusion frontier region includes: The predicted coupled risk increment of the reagent combination corresponding to each risk diffusion front region is obtained from the preset risk prediction module at a future time. The predicted coupled risk increment is generated based on the remaining amount decay curve and concentration change trend of the at least two reagents to be used. Based on the real-time field value within each risk diffusion frontier region and the predicted coupling risk increment, the rate of change of risk source terms in each risk diffusion frontier region is determined. The rate of change of risk source terms is used to characterize the driving strength of the newly added coupling risk on the evolution of the risk field. Based on the risk field change gradient within each risk diffusion frontier region and the spatial range pointed to by the risk field change gradient, the risk diffusion propagation change rate of each risk diffusion frontier region is determined. The risk diffusion propagation change rate is used to characterize the rate at which the risk spreads to the neighboring space. The risk source term change rate is superimposed and fused with the risk diffusion and propagation change rate to obtain the risk evolution rate corresponding to each risk diffusion frontier region.

9. The method according to claim 7, characterized in that, The process of dynamically generating a model based on the permission policy to comprehensively calculate the current risk status and future risk trends, outputting and executing a permission control policy bound to the current risk level, includes: The real-time field value and the risk evolution rate are fused together to obtain the predicted risk field value of each risk diffusion frontier region at a future preset time. The real-time field value and the predicted risk field value are compared with the multi-level safety threshold, and the current risk level and the predicted risk level are determined based on the comparison results. Based on the level transition relationship between the current risk level and the predicted risk level, the policy generation rule corresponding to the level transition relationship is called from the permission policy dynamic generation model to obtain a combination of permission control parameters including authentication factor combination method, reagent unlocking range and monitoring sampling frequency; The access control parameter combination is sent to the execution module, which then authenticates the current user according to the authentication factor combination, unlocks the electronic lock array of the hazardous chemical cabinet according to the reagent unlocking range, and dynamically adjusts the sampling frequency of the sensor network according to the monitoring sampling frequency.

10. A system for batch dispensing and access control of multiple reagents in a hazardous chemical cabinet, characterized in that, The system includes: The retrieval module is used to respond to a batch retrieval request containing at least two reagents to be retrieved, retrieve chemical kinetic parameters corresponding to the reagents to be retrieved from a pre-stored chemical characteristic parameter library, the chemical kinetic parameters including reaction activation energy and reaction heat, and apply a temperature parameter from real-time environmental monitoring data to the reaction activation energy to obtain a temperature-corrected activation energy parameter; The calculation module is used to calculate the coupling risk coefficient of the at least two reagents to be taken in real time based on the temperature-corrected activation energy parameter, the heat of reaction and the real-time environmental monitoring data, and to aggregate the calculation results of all reagent combinations to generate a multi-reagent coupling risk matrix. The multi-reagent coupling risk matrix is ​​used to quantitatively characterize the chemical incompatibility risk of all reagent combinations in the batch take-up list in the real-time environment. The input module is used to input the multi-reagent coupled risk matrix and the cabinet spatial location data obtained from the sensor network into the spatiotemporal risk field evolution model. The spatiotemporal risk field evolution model is used to iteratively solve the distribution and evolution trend of risk in the spatial and temporal dimensions, and output the four-dimensional spatiotemporal risk field under the current operation scenario. The execution module is used to obtain and execute an access control strategy bound to the current risk level based on the comparison results between the real-time field value of the four-dimensional spatiotemporal risk field and the multi-level security threshold. The access control strategy includes the linkage adjustment of the authentication factor combination method, reagent unlocking range and monitoring sampling frequency.

Images