An artificial intelligence driven catalytic industry control method and system

Abstract

This invention relates to the field of industrial control technology and discloses an artificial intelligence-driven method and system for catalytic industrial control. The method includes: calculating the characteristic levels of multiple bed spatial partitioning units and catalyst particles, and assigning corresponding operating condition benchmark values ​​to the bed spatial partitioning units based on the target reaction conditions; calculating the transfer and reaction characteristic quantities of the bed spatial partitioning units, splicing and reconstructing the bed state field and extracting the gradient distribution; mapping reaction types and performing risk coupling scoring on the bed spatial partitioning units in conjunction with the catalytic reaction path; calling the bed partition particle structure data to match the characteristic levels of the corresponding catalyst particles, performing differentiated catalyst particle configuration on the bed spatial partitioning units, and generating catalyst classification configuration data to guide industrial loading. This invention achieves fine modeling of the bed reaction environment, reaction type identification, and risk assessment, drives differentiated catalyst configuration, and improves the conversion rate and stability of the target reactants.

Description

Technical Field

[0001] This invention relates to the field of chemical catalytic synthesis, and more particularly to an artificial intelligence-driven industrial control method and system for catalysis. Background Technology

[0002] Industrial catalytic reactions are a core technology in modern petrochemical, coal chemical, fine chemical, and new material synthesis. Their control precision directly determines product selectivity, feedstock conversion rate, energy and material consumption, catalyst lifetime, and plant operational stability. Methacrylonitrile is an important chemical intermediate widely used in the synthesis of specialty resins, functional polymers, heat-resistant rubbers, and high-performance material monomers. With the development of the fine chemical and new materials industries, higher demands are placed on the selectivity, stability, and energy efficiency of methacrylonitrile production processes. Taking the conversion process of methacrylonitrile as an example, multi-component metal oxide catalysts are commonly used industrially to convert corresponding hydrocarbon feedstocks into methacrylonitrile through gas-phase ammonia oxidation. This type of reaction is strongly exothermic, with complex reaction pathways, numerous intermediates, and a high degree of coupling between catalyst surface reactions and gas-solid mass transfer processes. Therefore, catalyst performance and the reaction environment have a significant impact on the final product distribution.

[0003] In existing research, most work focuses on optimizing the catalyst materials themselves, such as improving the selectivity of the target product by adjusting the metal element composition, controlling the surface acid-base properties, optimizing the active phase structure, or introducing promoters. Meanwhile, some studies also improve reaction efficiency by optimizing macroscopic process conditions such as reaction temperature, feed ratio, and space velocity. For example, patent CN118718796A proposes a highly efficient and dust-free catalyst preparation device and method for ammonia oxime reaction. A liquid jet pump creates negative pressure to draw the catalyst from the container into a preparation tank for mixing. Combined with a stirrer and reflux pipeline, the catalyst slurry is circulated, reducing the risk of dust and minimizing manual labor. However, in the actual operation of industrial fixed-bed reactors, the catalyst is often packed in particulate form inside the bed. Its particle size distribution, morphology, and packing method create complex packing structures within the bed, affecting gas flow paths, local mass transfer rates, and the spatial distribution of exothermic reactions. Traditional methods and research cannot achieve differentiated and precise preparation for complex reaction environments.

[0004] In laboratory-scale studies, reactors are typically small, with relatively uniform catalyst loading and ideal gas flow distribution, thus the bed is often simplified as a homogeneous reaction system. However, under industrial-scale conditions, due to the inevitable differences in catalyst particle size distribution, fine-particle enrichment zones, coarse-particle channel zones, or locally densely packed zones can easily form during loading. These structural differences lead to localized mass transfer differences and temperature gradients within the bed at the millimeter to centimeter scale. For the highly exothermic gas-phase ammonia oxidation reaction, this inhomogeneity of the microscale reaction environment can be further amplified, resulting in uneven distribution of local temperature peaks and ammonia concentrations. This alters the reaction pathways at the catalyst active sites, making some regions more prone to generating byproducts such as methacrylamide or deeply oxidized products.

[0005] To address this problem, this invention proposes an artificial intelligence-driven method and system for controlling catalytic industrial reactions, achieving high efficiency and stability in the catalytic industrial reaction system. Summary of the Invention

[0006] This invention proposes an AI-driven industrial control method and system for catalysis. Addressing the problems of strong bed inhomogeneity, difficulty in quantifying mass and heat transfer coupling, and reliance on experience in catalyst configuration during existing catalytic reactions, step S1 utilizes multi-source detection and spatial partitioning mapping methods, combined with feature level calculation and operating condition baseline assignment, to solve the problems of inconsistent characterization of particle structure and reaction conditions, and the inability to quantify spatial distribution in traditional catalyst beds. Step S2, by constructing transfer and reaction feature quantities and performing bed state field splicing reconstruction and gradient extraction, solves the problems of difficulty in finely characterizing mass and heat transfer and reaction states within the bed, and the lack of continuous description. Step S3, combining microscale reaction environment distribution data with catalytic reaction path mapping and introducing a risk coupling scoring mechanism, solves the problems of difficulty in accurately identifying the main reaction and various side reactions, and the difficulty in quantifying reaction risks. Step S4, through a differentiated catalyst particle configuration strategy based on reaction type and risk level, solves the problems of existing catalyst loading relying on experience and lacking regionally targeted optimization, thereby achieving refined and adaptive optimization of catalyst configuration.

[0007] To achieve the above objectives, the present invention provides an artificial intelligence-driven catalytic industrial control method, comprising the following steps: S1: Collect raw data of catalyst particles and industrial bed reactor using multi-source detection methods, perform spatial partitioning mapping and feature level calculation on the raw data to obtain multiple bed spatial partitioning units and the feature level of catalyst particles, and assign corresponding operating condition benchmark values ​​to the bed spatial partitioning units in combination with the target reaction conditions to construct the bed partitioning particle structure data. S2: Based on the particle structure data of the bed partition, combined with the target reaction conditions, calculate the transmission and reaction characteristics of the bed spatial partition unit, splice and reconstruct the bed state field and extract the gradient distribution to obtain microscale reaction environment distribution data. S3: Based on the microscale reaction environment distribution data, the reaction type mapping and risk coupling score of the bed spatial partition unit are performed in combination with the catalytic reaction path to obtain the reaction type and risk coupling score of the bed spatial partition unit. S4: Based on the reaction type and risk coupling score of the bed space partitioning unit, call the bed partitioning particle structure data to match the characteristic level of the corresponding catalyst particles, and perform differentiated catalyst particle configuration on the bed space partitioning unit to generate catalyst classification configuration data to guide industrial loading.

[0008] As a further improvement of the present invention: Further, in step S1, raw data of catalyst particles and industrial bed reactor are collected using multi-source detection methods. Spatial partitioning and feature level calculation are performed on the raw data to obtain multiple bed spatial partitioning units and the feature levels of the catalyst particles, including: S11: Obtain catalyst particles, and use a multi-layer standard sieve method to sieve the catalyst particles to obtain a set of catalyst particles under multiple sieve aperture sizes. Randomly select catalyst particles from each set of sieve aperture sizes to obtain a sample of catalyst particles under each set of sieve aperture sizes. S12: The catalyst particle sample is scanned using a laser three-dimensional scanning device to extract the surface roughness, number of edges and corners, sphericity, and surface porosity of the catalyst particle sample as morphological data of the catalyst particle sample; The catalyst particle sample was analyzed using a low-temperature nitrogen adsorption method to obtain the specific surface area, total pore volume, and average pore size of the catalyst particle sample, which were used as the pore structure data of the catalyst particle sample. The static packing density, tapped packing density, and bed porosity of the catalyst particle sample were measured and used as the packing density data of the catalyst particle sample. S13: Use the morphology data, pore structure data, and bulk density data of the catalyst particle sample as the original data of the sieve pore size associated with the catalyst particle sample. S14: Standardize the original data of the sieve aperture particle size, and weight the standardized morphology data to obtain the morphology index score of the sieve aperture particle size; weight the standardized pore structure data to obtain the pore structure index score of the sieve aperture particle size; weight the standardized bulk density data to obtain the bulk density index score of the sieve aperture particle size. The morphology score, pore structure score, and bulk density score of the sieve pore size are used as the characteristic levels of the sieve pore size. S15: The industrial bed reactor is divided into multiple bed space partitioning units using a three-dimensional mesh partitioning method. The unique number, axial start and end coordinates, radial start and end coordinates, circumferential start and end coordinates, center point coordinates, theoretical volume, and distance from the heat exchange tubes of each bed space partitioning unit are recorded as the spatial partitioning mapping result of the industrial bed reactor.

[0009] Furthermore, step S1, which assigns baseline values ​​to the bed spatial partitioning units based on the target reaction conditions to construct the bed partitioning particle structure data, also includes: S16: Obtain the target reaction conditions in the catalytic process of the target reactant, wherein the target reaction conditions include hydrocarbon feed concentration, ammonia feed concentration, oxygen feed concentration, feed temperature, gas reactant flow rate and the maximum heat release per unit volume of catalyst bed, and summarize the hydrocarbon feed concentration, ammonia feed concentration and oxygen feed concentration as the total inlet flow rate; S17: Based on the distance between the bed space partitioning unit and the heat exchange tube, and the characteristic grade of the sieve particle size, combined with the total inlet flow rate and the maximum heat release per unit volume of catalyst bed, the operating condition reference value between the bed space partitioning unit and the catalyst particles associated with the sieve particle size is calculated, wherein the operating condition reference value includes the feed flow rate reference value and the heat release load reference value. Specifically, the formula for calculating the operating baseline value between the bed space partitioning unit and the catalyst particles associated with the sieve aperture size is as follows: ; ; ; in, Indicates the spatial partitioning unit of the nth bed layer and the... The operating baseline value between catalyst particles associated with the sieve aperture size. These represent the reference values ​​for the operating conditions, respectively. The feed throughput benchmark value and the exothermic load benchmark value are included. The number of groups representing the sieve aperture size indicates the total number of bed space partitioning units; This represents the total inlet flow rate determined based on the target reaction conditions. This represents the radial projection area of ​​the nth bed space partition unit in the industrial bed reactor. This represents the total cross-sectional area of ​​the industrial bed reactor in the radial direction (calculated from the inner diameter of the industrial bed reactor, using the following method). (the square of the inner diameter) The characteristic grade of the sieve aperture size in the group is represented by , where , They represent the number respectively. The scores for morphology, pore structure, and bulk density of the sieve particles were calculated. This indicates that packing density affects the weight. This indicates that the shape affects the weight. This indicates the influence of hole structure on weights; This indicates the maximum heat release from the reaction that a unit volume of catalyst bed can withstand. This represents the spatial location correction factor for the nth bed-level spatial partition unit. This represents the axial attenuation correction factor for the nth bed space partition unit; S18: The spatial partitioning mapping result of the industrial bed reactor, the operating condition benchmark value between the bed spatial partitioning unit and the catalyst particles associated with the sieve aperture size, and the characteristic level of the sieve aperture size are used as the bed partitioning particle structure data.

[0010] Furthermore, in step S2, the transfer and reaction characteristics of the bed space partitioning unit are calculated in conjunction with the target reaction conditions, including: S21: Based on the particle structure data of the bed partition and the target reaction conditions, calculate the local gas flow obstruction intensity, local external diffusion transfer intensity and local internal diffusion transfer intensity of the bed space partition unit under the action of catalyst particles associated with the sieve aperture size, and use the local gas flow obstruction intensity, local external diffusion transfer intensity and local internal diffusion transfer intensity as the transfer characteristic quantities of the bed space partition unit under the action of catalyst particles associated with the sieve aperture size; S22: Calculate the local heat accumulation intensity, local ammonia dissipation rate, and unit intermediate residence intensity of the bed space partitioning unit under the action of catalyst particles associated with the sieve aperture size. Use the local heat accumulation intensity, local ammonia dissipation rate, and unit intermediate residence intensity as the reaction characteristic quantities of the bed space partitioning unit under the action of catalyst particles associated with the sieve aperture size.

[0011] Furthermore, step S2 involves stitching together and reconstructing the bed state field and extracting the gradient distribution to obtain microscale reaction environment distribution data, which also includes: S23: Based on the aforementioned transfer and reaction characteristics, calculate the physical state quantities of the bed space partitioning unit under the action of catalyst particles associated with different sieve aperture sizes, wherein the physical state quantities include unit temperature value, unit ammonia concentration value, unit oxygen concentration value, and unit intermediate residence intensity. S24: Calculate the splicing weight between the bed space partitioning unit and the adjacent unit, and smoothly reconstruct the physical state quantity of the bed space partitioning unit to obtain the spatially continuous physical state quantity as the bed state field. S25: Based on the bed state field, extract the gradient difference index of the bed spatial partition unit as microscale reaction environment distribution data, wherein the gradient difference index includes temperature gradient index, ammonia concentration gradient index, oxygen concentration gradient index and intermediate gradient index.

[0012] Furthermore, in step S3, the reaction type mapping and risk coupling scoring of the bed spatial partitioning units are combined with the catalytic reaction pathway, including: S31: Based on the microscale reaction environment distribution data, extract the gradient difference index of the bed spatial partition unit, and combine it with the catalytic reaction path to generate the reaction type of the bed spatial partition unit, wherein the reaction type includes the main reaction, amide side reaction and deep oxidation side reaction; S32: Based on the reaction type of the bed spatial partitioning unit, calculate the mean gradient difference index of the bed spatial partitioning unit under the action of catalyst particles associated with the U-group sieve pore size. Apply a risk coupling weighting to the mean gradient difference index of the bed spatial partitioning unit to obtain the risk coupling score of the bed spatial partitioning unit. This indicates the number of groups representing the sieve aperture size.

[0013] Furthermore, the formula for risk-coupled weighting of the gradient difference index of the bed space partitioning unit in step S32 is as follows: ; in, This represents the risk coupling score of the nth bed-level spatial partition unit. , This indicates the total number of spatial partitioning units in the bed layer. This represents the mean gradient difference index of the nth bed-layer spatial partition unit. The values ​​represent the mean values ​​of the temperature gradient index, the ammonia concentration gradient index, the oxygen concentration gradient index, and the intermediate gradient index, respectively. This indicates the reaction type of the nth bed space partitioning unit.

[0014] Further, in step S4, the bed partition particle structure data is called to match the characteristic level of the corresponding catalyst particles, and differentiated catalyst particle configuration is performed on the bed spatial partition unit, including: S41: Based on the risk coupling score of the bed space partitioning unit, the bed space partitioning unit is divided into high-risk unit, medium-risk unit and low-risk unit; S42: Based on the risk unit type and reaction type of the bed space partitioning unit, match catalyst particles with sieve pore size corresponding to the characteristic level as catalyst classification configuration data, and configure the catalyst particles matched by the bed space partitioning unit in the bed space partitioning unit.

[0015] The present invention also provides an artificial intelligence-driven catalytic industrial control system, which includes a catalyst particle storage device and an industrial bed reactor. The catalyst particle storage device is used to store catalyst particles of multiple sieve sizes, and the industrial bed reactor is used for raw material conversion and target reactant generation, and provides a stable reaction space and operating environment for the catalytic reaction, so as to realize the artificial intelligence-driven catalytic industrial control method described above.

[0016] Compared with existing technologies, this invention proposes an artificial intelligence-driven control method and system for the catalytic industry, which has the following beneficial effects: First, this invention constructs a partitioned chemical condition benchmark model for the catalytic process of target reactants by introducing multi-source operating parameters and a bed spatial partitioning mapping mechanism, achieving synergistic quantitative characterization of feed throughput and exothermic carrying capacity. Specifically, by coupling and correcting the characteristic grade of sieve particle size with spatial position parameters (such as heat transfer distance and axial layer number), the adaptability of catalyst particles in different bed spatial partitioning units can be finely adjusted according to dynamic changes in the reaction environment, thereby effectively improving the degree of mass and heat transfer matching within the bed. Furthermore, by calculating the operating condition benchmark value by unit, local overheating or uneven reaction problems are avoided, enhancing the stability and safety of the reaction process, thus providing a data-driven basis for subsequent catalyst staged loading, significantly improving catalytic efficiency and product selectivity, and reducing energy consumption and operational risks.

[0017] Meanwhile, this invention maps transport and reaction characteristics to key state variables such as temperature, ammonia concentration, oxygen concentration, and intermediate residence intensity, and combines a neighborhood weighting strategy to eliminate abrupt errors caused by partitioning and discretization, thus ensuring good continuity and stability of the bed state field in both the axial and radial directions. By constructing a gradient difference index, key risk areas such as local temperature rise anomalies, uneven reactant consumption, and intermediate accumulation can be quantitatively identified, providing high-resolution evidence for reaction path analysis and risk coupling assessment. This enhances the analytical capability for complex heterogeneous reaction fields, helps improve the targeting and precision of catalyst configuration decisions, optimizes overall reaction homogeneity, and reduces the risk of local runaway. Attached Figure Description

[0018] Figure 1 This is a schematic flowchart of an artificial intelligence-driven catalytic industrial control method provided in an embodiment of the present invention.

[0019] Figure 2 This is a structural diagram of an industrial bed reactor provided in one embodiment of the present invention.

[0020] Figure 3 This is an experimental comparison diagram provided for one embodiment of the present invention. Detailed Implementation

[0021] The realization of the objectives, functional characteristics, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0022] This invention provides an artificial intelligence-driven catalytic industrial control method and system. The executing entity of the artificial intelligence-driven catalytic industrial control method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this invention: a server, a terminal, etc. In other words, the artificial intelligence-driven catalytic industrial control method can be executed by software or hardware installed on a terminal device or a server device, and the software may be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0023] Reference Figure 1 Embodiment 1 of the present invention is as follows: An artificial intelligence-driven method for controlling catalysis in an industrial setting includes the following steps: S1: Collect raw data of catalyst particles and industrial bed reactor using multi-source detection methods, perform spatial partitioning mapping and feature level calculation on the raw data to obtain multiple bed spatial partitioning units and the feature level of catalyst particles, and assign corresponding operating condition benchmark values ​​to the bed spatial partitioning units in combination with the target reaction conditions to construct the bed partitioning particle structure data.

[0024] Specifically, in step S1, raw data of catalyst particles and industrial bed reactor are collected using multi-source detection methods. Spatial partitioning and feature level calculation are then performed on the raw data to obtain multiple bed spatial partitioning units and the feature levels of the catalyst particles, including: S11: Obtain catalyst particles, and use a multi-layer standard sieve method to sieve the catalyst particles to obtain a set of catalyst particles under multiple sieve aperture sizes. Randomly select catalyst particles from each set of sieve aperture sizes to obtain a sample of catalyst particles under each set of sieve aperture sizes. Specifically, an 8-layer standard sieve (with sieve aperture sizes of 1.2mm, 1.5mm, 1.8mm, 2.1mm, 2.4mm, 2.7mm, 3.0mm, and 3.3mm respectively) was used for sieving to obtain a set of catalyst particles with different sieve aperture sizes. 150 catalyst particles were randomly selected from the set of catalyst particles as a catalyst particle sample. S12: The catalyst particle sample is scanned using a laser three-dimensional scanning device to extract the surface roughness, number of edges and corners, sphericity, and surface porosity of the catalyst particle sample as morphological data of the catalyst particle sample; The catalyst particle sample was analyzed using a low-temperature nitrogen adsorption method to obtain the specific surface area, total pore volume, and average pore size of the catalyst particle sample, which were used as the pore structure data of the catalyst particle sample. The static packing density, tapped packing density, and bed porosity of the catalyst particle sample were measured and used as the packing density data of the catalyst particle sample. S13: Use the morphology data, pore structure data, and bulk density data of the catalyst particle sample as the original data of the sieve pore size associated with the catalyst particle sample. S14: Standardize the original data of the sieve aperture particle size, and weight the standardized morphology data to obtain the morphology index score of the sieve aperture particle size; weight the standardized pore structure data to obtain the pore structure index score of the sieve aperture particle size; weight the standardized bulk density data to obtain the bulk density index score of the sieve aperture particle size. The morphology score, pore structure score, and bulk density score of the sieve pore size are used as the characteristic levels of the sieve pore size. Specifically, the scores for morphology, pore structure, and bulk density are all between 0 and 1. A higher morphology score indicates a rougher and more irregular surface of the catalyst particles corresponding to the sieve size. A higher pore structure score indicates better connectivity of the internal channels and a more developed pore network structure of the catalyst particles corresponding to the sieve size. A higher bulk density score indicates a denser packing and lower porosity of the catalyst particles corresponding to the sieve size within the bed space partitioning unit. Specifically, artificial intelligence technologies (such as neural network models or reinforcement learning models) are used to generate characteristic grades of sieve aperture size; S15: The industrial bed reactor is divided into multiple bed space partitioning units using a three-dimensional mesh partitioning method. The unique number, axial start and end coordinates, radial start and end coordinates, circumferential start and end coordinates, center point coordinates, theoretical volume, and distance from the heat exchange tubes of each bed space partitioning unit are recorded as the spatial partitioning mapping result of the industrial bed reactor.

[0025] Step S1, which assigns baseline values ​​to the bed spatial partitioning units based on the target reaction conditions to construct the bed partitioning particle structure data, also includes: S16: Obtain the target reaction conditions in the catalytic process of the target reactant, wherein the target reaction conditions include hydrocarbon feed concentration, ammonia feed concentration, oxygen feed concentration, feed temperature, gas reactant flow rate and the maximum heat release per unit volume of catalyst bed, and summarize the hydrocarbon feed concentration, ammonia feed concentration and oxygen feed concentration as the total inlet flow rate; In one embodiment of the present invention, methacrylonitrile is used as the target reactant; S17: Based on the distance between the bed space partitioning unit and the heat exchange tube, and the characteristic grade of the sieve particle size, combined with the total inlet flow rate and the maximum heat release per unit volume of catalyst bed, the operating condition reference value between the bed space partitioning unit and the catalyst particles associated with the sieve particle size is calculated, wherein the operating condition reference value includes the feed flow rate reference value and the heat release load reference value. Specifically, the formula for calculating the operating baseline value between the bed space partitioning unit and the catalyst particles associated with the sieve aperture size is as follows: ; ; ; in, This represents the baseline operating condition between the nth bed spatial partition unit and the catalyst particles associated with the sieve aperture size group. These represent the reference values ​​for the operating conditions, respectively. The feed throughput benchmark value and the exothermic load benchmark value are included. Indicates the number of groups representing the sieve aperture size. This indicates the total number of spatial partitioning units in the bed layer; This represents the total inlet flow rate determined based on the target reaction conditions. This represents the radial projection area of ​​the nth bed space partition unit in the industrial bed reactor. This represents the total cross-sectional area of ​​the industrial bed reactor in the radial direction (calculated from the inner diameter of the industrial bed reactor, using the following method). (the square of the inner diameter) Indicates the first Characteristic grades of sieve aperture size, among which , The scores for morphology, pore structure, and bulk density of the sieve particles in the i-th group are respectively represented. This indicates that packing density affects the weight. This indicates that the shape affects the weight. This indicates that the hole structure affects the weight; the default setting is used. The values ​​are 0.3, 0.4, and 0.3 respectively. This indicates the maximum heat release from the reaction that a unit volume of catalyst bed can withstand. This represents the spatial location correction factor for the nth bed-level spatial partition unit. This represents the axial attenuation correction factor for the nth bed space partition unit; Specifically, the spatial position correction coefficient The correlation between the distance to the nth bed space partition unit and the heat exchange tube is negative; if the distance is less than 50 mm, the correlation is 1.3; if the distance is between 50 and 100 mm, the correlation is... The value is 1.2. If the distance is between 100-200 mm, then... The value is 1.1. If the distance is greater than 200 mm, then... =1; Based on the axial layer number of the nth bed space partitioning unit, determine the axial attenuation correction factor of the nth bed space partitioning unit. If the axial layer number of the nth bed space partition unit is the first 1 / 3, then If the axial layer number of the nth bed space partition unit is the last 1 / 3, then... The value is 0.9. If the axial layer number of the nth bed-layer spatial partition unit is otherwise, then... It is 0.98; S18: The spatial partitioning mapping result of the industrial bed reactor, the operating condition benchmark value between the bed spatial partitioning unit and the catalyst particles associated with the sieve aperture size, and the characteristic level of the sieve aperture size are used as the bed partitioning particle structure data.

[0026] S2: Based on the particle structure data of the bed partitions and the target reaction conditions, calculate the transmission and reaction characteristics of the bed spatial partition units, reconstruct the bed state field and extract the gradient distribution to obtain microscale reaction environment distribution data.

[0027] In step S2, the transfer and reaction characteristics of the bed space partitioning unit are calculated in conjunction with the target reaction conditions, including: S21: Based on the particle structure data of the bed partition and the target reaction conditions, calculate the local gas flow obstruction intensity, local external diffusion transfer intensity and local internal diffusion transfer intensity of the bed space partition unit under the action of catalyst particles associated with the sieve aperture size, and use the local gas flow obstruction intensity, local external diffusion transfer intensity and local internal diffusion transfer intensity as the transfer characteristic quantities of the bed space partition unit under the action of catalyst particles associated with the sieve aperture size; Specifically, the calculation formulas for the local gas passage obstruction intensity, local external diffusion transfer intensity, and local internal diffusion transfer intensity of the bed space partitioning unit under the action of catalyst particles associated with the sieve aperture size are as follows: ; ; ; ; in, This represents the transport characteristic quantity of the nth bed spatial partition unit under the action of catalyst particles associated with the sieve aperture size of the nth group. These represent the transmitted feature quantities respectively. The local gas passage obstruction intensity, the local external diffusion transmission intensity, and the local internal diffusion transmission intensity; This represents the topographic hindrance weighting coefficient (default setting is 0.4). This represents the accumulation resistance weighting coefficient (default setting is 0.5). This represents the gas clearance weighting coefficient (default setting is 0.25). Represents N bed-level spatial partitioning units and the first The average feed throughput baseline value among catalyst particles associated with the sieve aperture size; This represents the flow rate enhancement factor (default setting is 0.35). This represents the morphology diffusion coefficient (default setting is 0.3). This represents the pore structure dominance factor (default setting is 0.5). This indicates the feed temperature in the target reaction condition, and this indicates the reference temperature (set to 400 Kelvin for the methacrylonitrile synthesis condition). This represents the temperature diffusion promotion coefficient (default setting is 0.2). This represents the stacking inhibition coefficient (default setting is 0.25). Specifically, in calculating the local gas flow resistance strength, surface roughness and dense packing enhance resistance, while high feed throughput weakens resistance; in calculating the local external diffusion transfer strength, the faster the flow rate and the rougher the catalyst particle surface, the higher the external diffusion efficiency; in calculating the local internal diffusion transfer strength, good pore connectivity and high feed temperature enhance internal diffusion, while dense packing hinders internal diffusion. S22: Calculate the local heat accumulation intensity, local ammonia dissipation rate, and unit intermediate residence intensity of the bed space partitioning unit under the action of catalyst particles associated with the sieve aperture size. Use the local heat accumulation intensity, local ammonia dissipation rate, and unit intermediate residence intensity as the reaction characteristic quantities of the bed space partitioning unit under the action of catalyst particles associated with the sieve aperture size.

[0028] Specifically, the formulas for calculating the local heat accumulation intensity, local ammonia dissipation rate, and intermediate residence intensity of the bed space partitioning unit under the action of catalyst particles associated with the sieve aperture size are as follows: ; ; ; ; in, This indicates that the nth bed-layer spatial partition unit is in the nth... The reaction characteristic vector associated with the particle size of the sieve group and the action of the catalyst particles. These represent the reaction characteristic vectors respectively. The intensity of local heat accumulation, the rate of local ammonia dissipation, and the residence intensity of unit intermediates; This represents the heat load factor (default setting is 0.45). This represents the heat storage resistance coefficient (default setting is 0.35). This indicates the ammonia feed concentration under the target reaction conditions. This represents the concentration-driven coefficient (default setting is 0.35). This represents the external diffusion driving coefficient (default setting is 0.4). This represents the internal diffusion driving coefficient (default setting is 0.4). This represents the overheating suppression coefficient (default setting is 0.3). This represents the dwell time resistance coefficient (default setting is 0.45). This represents the flux retention inhibition coefficient (default setting is 0.3). This indicates the duct unblocking coefficient (default setting is 0.28). Specifically, a higher local gas flow resistance intensity indicates a greater flow obstruction of gas within the bed space partitioning unit; a higher local external diffusion transfer intensity indicates a higher efficiency of reactants (hydrocarbons, ammonia, oxygen) diffusion to the catalyst particle surface; and a higher local internal diffusion transfer intensity indicates a higher efficiency of reactants diffusion from the catalyst particle surface to the active sites inside the particles. A higher local heat accumulation intensity indicates a more significant accumulation of exothermic reaction within the bed space partition unit, and a greater risk of local overheating. A higher local ammonia dissipation rate indicates a faster consumption rate of ammonia in the bed space partition unit. A higher intermediate residence intensity indicates a longer residence time of intermediates generated during the reaction within the bed space partition unit. These intermediates are those generated during the reaction of methacrylonitrile, including imine intermediates, oxygen-containing intermediates (aldehydes, alcohols), etc.

[0029] Specifically, this invention achieves a refined characterization of the entire process of mass transfer, reaction, and heat release within the spatially partitioned units of the bed by constructing a hierarchical coupling computational mechanism for transport and reaction characteristics. Specifically, by fusing the characteristic levels corresponding to the sieve aperture size with the target reaction conditions, gas flow obstruction, internal and external diffusion efficiency, and heat accumulation behavior can be quantitatively analyzed, thereby accurately reflecting the reactivity and risk status of different regions. Furthermore, by introducing ammonia dissipation rate and intermediate residence intensity indices, regions with excessively rapid local reactions or intermediate accumulation can be effectively identified, avoiding the problem of enhanced side reactions. This provides high-resolution data support for subsequent reaction type determination and differentiated catalyst particle configuration, helping to improve the overall reaction uniformity, thermal management capability, and product yield of the bed, and significantly reducing the risk of local overheating and deactivation.

[0030] Step S2 involves stitching together and reconstructing the bed state field and extracting the gradient distribution to obtain microscale reaction environment distribution data. This also includes: S23: Based on the aforementioned transfer and reaction characteristics, calculate the physical state quantities of the bed space partitioning unit under the action of catalyst particles associated with different sieve aperture sizes, wherein the physical state quantities include unit temperature value, unit ammonia concentration value, unit oxygen concentration value, and unit intermediate residence intensity. Specifically, the nth bed-layer spatial partitioning unit is in the... The formula for calculating the physical state parameters of the catalyst particles associated with the sieve pore size is as follows: ; ; ; ; ; in, This indicates that the nth bed-layer spatial partition unit is in the nth... The physical state of the catalyst particles as a function of the sieve pore size. These represent the physical state quantities respectively. The unit temperature value, unit ammonia concentration value, unit oxygen concentration value, and unit intermediate residence intensity are included. This represents the heat-to-temperature-rise conversion coefficient (default value is 60 Kelvin·m³ / kW). This indicates that the additional temperature rise is inhibited (value is 9 Kelvin). This represents the ammonia dissipation-concentration decay coefficient (default value is 0.6). This indicates the oxygen feed concentration under the target reaction conditions. This represents the oxygen supply inhibition coefficient (default value is 0.2). This represents the heat-driven oxygen consumption coefficient (default value is 0.45). S24: Calculate the splicing weight between the bed space partitioning unit and the adjacent unit, and smoothly reconstruct the physical state quantity of the bed space partitioning unit to obtain the spatially continuous physical state quantity as the bed state field. As an embodiment of the present invention, if two bed space partitioning units have adjacent regions, then the two bed space partitioning units are adjacent units to each other. For the nth bed space partitioning unit in the nth... The physical state parameters of catalyst particles under the influence of sieve pore size, taking unit temperature as an example. The formula for smooth reconstruction is: ; ; in, Indicates unit temperature value The smooth reconstruction result Let e ​​represent the set of adjacent units of the nth bed-level spatial partition unit, and let e represent the set of adjacent units. Adjacent units in This represents the unit temperature value of adjacent unit e. This represents the splicing weight between the nth bed-level spatial partition unit and its adjacent unit e. This represents the axial distance between the nth bed-level spatial partition unit and its adjacent unit e. This represents the radial distance between the nth bed-level spatial partition unit and its adjacent unit e; S25: Based on the bed state field, extract the gradient difference index of the bed spatial partition unit as microscale reaction environment distribution data, wherein the gradient difference index includes temperature gradient index, ammonia concentration gradient index, oxygen concentration gradient index and intermediate gradient index.

[0031] Specifically, by obtaining the spatially continuous physical state quantities of the bed space partitioning unit in the axially nearest adjacent unit, and calculating the rate of change of the spatially continuous physical state quantities between the bed space partitioning unit and the axially nearest adjacent unit, the gradient difference index of the bed space partitioning unit is used. The formula for calculating the gradient difference index is as follows: ; ; ; ; ; in, This indicates that the nth bed-layer spatial partition unit is in the nth... Gradient difference index of catalyst particle action associated with sieve pore size The gradient difference indices are represented in order. The temperature gradient index, ammonia concentration gradient index, oxygen concentration gradient index, and intermediate gradient index are included. The nth bed-layer spatial partitioning unit is in the nth... The smoothing results of unit temperature, unit ammonia concentration, unit oxygen concentration, and unit intermediate residence intensity under the action of catalyst particles related to the group sieve pore size. This represents the radial distance between the nth bed space partition unit and the nearest bed space partition unit. These are, in order, the bed space partitioning unit that is radially closest to the nth bed space partitioning unit at the nth bed space partitioning unit. The smoothing results of unit temperature, unit ammonia concentration, unit oxygen concentration, and unit intermediate residence intensity under the action of catalyst particles related to the group sieve pore size.

[0032] S3: Based on the microscale reaction environment distribution data, the reaction type mapping and risk coupling score of the bed spatial partition unit are performed in combination with the catalytic reaction path to obtain the reaction type and risk coupling score of the bed spatial partition unit.

[0033] In step S3, the reaction type mapping and risk coupling scoring of the bed spatial partitioning units are performed in conjunction with the catalytic reaction pathway, including: S31: Based on the microscale reaction environment distribution data, extract the gradient difference index of the bed spatial partition unit, and combine it with the catalytic reaction path to generate the reaction type of the bed spatial partition unit, wherein the reaction type includes the main reaction, amide side reaction and deep oxidation side reaction; Specifically, based on the catalytic reaction pathway of the methacrylonitrile, including the main reaction (hydrocarbon + ammonia + oxygen → methacrylonitrile), side reaction 1 (amide formation: intermediate + ammonia → methacrylamide) and side reaction 2 (deep oxidation: intermediate + oxygen → carbon monoxide or carbon dioxide), the reaction type of the bed space partitioning unit is generated by combining the gradient difference index. As an embodiment of the present invention, the process for generating the reaction type of the bed space partitioning unit is as follows: S311: Calculate the average gradient difference index of the bed space partition unit under the action of catalyst particles associated with the U group sieve particle size, wherein the average gradient difference index includes the average temperature gradient index, the average ammonia concentration gradient index, the average oxygen concentration gradient index, and the average intermediate gradient index. S312: If the average value of the temperature gradient index is less than 0.4, the average value of the ammonia concentration gradient index is less than 0.3, the average value of the oxygen concentration gradient index is less than 0.3, and the average value of the intermediate gradient index is less than 0.4, it indicates that the gradients of various indicators in the reaction process of the bed space partitioning unit are gentle, the reaction environment is stable, which is conducive to the main reaction, and the reaction type is set as the main reaction. S313: If the average value of the ammonia concentration gradient index is greater than 0.3 and the average value of the intermediate gradient index is greater than 0.4, it indicates that the ammonia concentration gradient and the intermediate gradient are large, and the intermediate and ammonia are prone to side reactions. Set the reaction type to amide side reaction. S314: Set the reaction type of the bed space flipping unit that does not satisfy the main reaction or amide side reaction as deep oxidation side reaction; S32: Based on the reaction type of the bed spatial partitioning unit, calculate the mean gradient difference index of the bed spatial partitioning unit under the action of catalyst particles associated with the U-group sieve pore size. Apply a risk coupling weighting to the mean gradient difference index of the bed spatial partitioning unit to obtain the risk coupling score of the bed spatial partitioning unit. This indicates the number of groups representing the sieve aperture size.

[0034] The formula for risk-coupled weighting of the gradient difference index of the bed space partitioning unit in step S32 is as follows: ; in, This represents the risk coupling score of the nth bed-level spatial partition unit. , This indicates the total number of spatial partitioning units in the bed layer. This represents the mean gradient difference index of the nth bed-layer spatial partition unit. The values ​​represent the mean values ​​of the temperature gradient index, the ammonia concentration gradient index, the oxygen concentration gradient index, and the intermediate gradient index, respectively. This indicates the reaction type of the nth bed space partitioning unit.

[0035] It should be noted that this invention establishes an adaptive reaction type determination mechanism for bed spatial partitioning units by fusing gradient difference indicators in microscale reaction environment distribution data with catalytic reaction pathways, achieving accurate differentiation between main reactions and various side reactions. Specifically, this invention utilizes a collaborative threshold discrimination method to effectively characterize the stability of the reaction environment and the material coupling state, thereby avoiding misjudgment problems caused by traditional single-indicator judgment. At the same time, it introduces a risk coupling weighted calculation method based on reaction type, and assigns differentiated weights to different gradient indicators, enabling risk assessment to be dynamically adjusted according to the characteristics of main reactions, amide side reactions, and deep oxidation side reactions. This not only accurately identifies high-risk areas and areas prone to side reactions, but also provides a quantitative decision-making basis for subsequent catalyst particle classification and configuration, significantly improving reaction selectivity, stability, and overall safety.

[0036] S4: Based on the reaction type and risk coupling score of the bed space partitioning unit, call the bed partitioning particle structure data to match the characteristic level of the corresponding catalyst particles, and perform differentiated catalyst particle configuration on the bed space partitioning unit to generate catalyst classification configuration data to guide industrial loading.

[0037] Specifically, step S4 involves calling the bed partition particle structure data to match the characteristic levels of the corresponding catalyst particles and performing differentiated catalyst particle configuration on the bed spatial partition units, including: S41: Based on the risk coupling score of the bed space partitioning unit, the bed space partitioning unit is divided into high-risk unit, medium-risk unit and low-risk unit; As an embodiment of the present invention, if risk coupling scoring Then the nth bed-level spatial partition unit is classified as a high-risk unit, if the risk coupling score Then the nth bed-level spatial partition unit is classified as a medium-risk unit, if the risk coupling score Then the nth bed-level spatial partition unit is divided into a low-risk unit; S42: Based on the risk unit type and reaction type of the bed space partitioning unit, match catalyst particles with sieve pore size corresponding to the characteristic level as catalyst classification configuration data, and configure the catalyst particles matched by the bed space partitioning unit in the bed space partitioning unit.

[0038] In another embodiment of the present invention, for the main reaction (low-risk unit): catalyst particles with sieve pore sizes matching medium-low morphology index scores (0.3-0.5), medium-high porosity index scores (0.6-0.8), and medium bulk density index scores (0.4-0.6) are used to ensure the stable progress of the main reaction; for the main reaction (medium / high-risk unit): catalyst particles with sieve pore sizes matching medium morphology index scores (0.4-0.6), high porosity index scores (0.7-0.9), and medium bulk density index scores (0.4-0.6) are used to improve mass transfer efficiency and suppress minor side reactions; For amide side reactions (medium / high risk units): catalyst particles with sieve sizes matching low morphology scores (0.2-0.4), high porosity scores (0.7-0.9), and low bulk density scores (0.2-0.4) are used to reduce intermediate retention and improve ammonia diffusion efficiency. For deep oxidation side reactions (medium / high risk): catalyst particles with sieve sizes matching low morphology scores (0.2-0.4), medium-high porosity scores (0.6-0.8), and low bulk density scores (0.2-0.4) are used to enhance heat removal and reduce local oxygen enrichment.

[0039] It should be noted that this invention achieves differentiated and precise configuration of catalyst particles in the bed space through a two-dimensional matching mechanism based on risk coupling score and reaction type. It can dynamically optimize the combination of morphology, pore structure and packing density parameters for different reaction regions, select the optimal catalyst particles, effectively suppress side reactions, reduce the risk of local overheating, and improve mass transfer efficiency and reaction uniformity, thereby significantly improving the selectivity of methacrylonitrile formation and overall operational stability.

[0040] Example 2: As another embodiment of the present invention, this embodiment provides an artificial intelligence-driven catalytic industrial control system. The catalytic industrial control system includes a catalyst particle storage device and an industrial bed reactor. The catalyst particle storage device is used to store catalyst particles of multiple sieve sizes. The industrial bed reactor is used for raw material conversion and target reactant generation, and provides a stable reaction space and operating environment for the catalytic reaction, so as to realize the artificial intelligence-driven catalytic industrial control method as described in Example 1.

[0041] The industrial bed reactor (adapted to the methacrylonitrile synthesis process) is the core equipment for realizing raw material conversion and product generation. Its core function is to provide a stable reaction space and operating environment for the catalytic reaction. Specifically, it carries the catalyst particles and achieves their orderly loading, ensuring sufficient contact between the reactant gases (hydrocarbons, ammonia, oxygen, etc.) and the catalyst particles, and providing suitable temperature, pressure, and mass transfer conditions for the efficient execution of the main methacrylonitrile reaction, as described above. Figure 2The diagram shows the structure of an industrial bed reactor. Taking an industrial bed reactor with an inner diameter of 2200 mm and a total bed height of 4800 mm (including an 800 mm pretreatment section, a 3200 mm main reaction section, and an 800 mm posttreatment section) as an example, the partitioning process using a three-dimensional mesh method is as follows: axially, it is divided into 80 layers at a height of 60 mm (including 13 layers of pretreatment section, 53 layers of main reaction section, and 14 layers of posttreatment section); radially, it is divided into 11 concentric rings at a width of 80 mm; and circumferentially, it is divided into 6 sectors at 60° intervals. After removing the space occupied by heat exchange tubes, 5280 bed space partitioning units are finally formed, and each unit is assigned a unique number (bed section number + axial layer number + radial ring number + circumferential sector number).

[0042] Example 3 This invention uses a 2200mm inner diameter industrial bed reactor as the experimental platform to compare the operational effects of the traditional empirical catalyst preparation method with the methacrylonitrile catalytic optimization method described in this invention. The experiment maintained consistent operating parameters such as feed flow rate, temperature, and pressure, and statistics were collected after 72 hours of continuous operation. (Refer to...) Figure 3 The experimental comparison diagram shows that, under the methacrylonitrile catalytic configuration optimization method described in this invention, the conversion rate of methacrylonitrile is increased by 3.2%, the formation rate of methacrylamide is reduced by 68%, and the CO... x The formation rate of (mainly carbon monoxide and carbon dioxide) was reduced by 55%, the bed hot spot temperature fluctuation decreased from ±8 Kelvin to ±2 Kelvin, ammonia consumption decreased by 4.1 kg (per ton of methacrylonitrile produced), and the catalyst single-pass operation cycle was extended by 28 days. The results indicate that this invention, through precise bed partitioning catalytic configuration, effectively suppresses side reactions, stabilizes the reaction environment, and significantly improves process economy and stability.

[0043] It should be noted that the terms "comprising," "including," or any other variations thereof used herein are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0044] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0045] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. An artificial intelligence-driven method for controlling catalytic industry, characterized in that, The method includes: S1: Collect raw data of catalyst particles and industrial bed reactor using multi-source detection methods, perform spatial partitioning mapping and feature level calculation on the raw data to obtain multiple bed spatial partitioning units and the feature level of catalyst particles, and assign corresponding operating condition benchmark values ​​to the bed spatial partitioning units in combination with the target reaction conditions to construct the bed partitioning particle structure data. S2: Based on the particle structure data of the bed partition, combined with the target reaction conditions, calculate the transmission and reaction characteristics of the bed spatial partition unit, splice and reconstruct the bed state field and extract the gradient distribution to obtain microscale reaction environment distribution data. S3: Based on the microscale reaction environment distribution data, the reaction type mapping and risk coupling score of the bed spatial partition unit are performed in combination with the catalytic reaction path to obtain the reaction type and risk coupling score of the bed spatial partition unit. S4: Based on the reaction type and risk coupling score of the bed space partitioning unit, call the bed partitioning particle structure data to match the characteristic level of the corresponding catalyst particles, and perform differentiated catalyst particle configuration on the bed space partitioning unit to generate catalyst classification configuration data to guide industrial loading.

2. The artificial intelligence-driven catalytic industrial control method as described in claim 1, characterized in that, In step S1, multi-source detection methods are used to collect raw data of catalyst particles and industrial bed reactors. Spatial partitioning and feature level calculations are then performed on the raw data to obtain multiple bed spatial partitioning units and the feature levels of the catalyst particles, including: S11: Obtain catalyst particles, and use a multi-layer standard sieve method to sieve the catalyst particles to obtain a set of catalyst particles under multiple sieve aperture sizes. Randomly select catalyst particles from each set of sieve aperture sizes to obtain a sample of catalyst particles under each set of sieve aperture sizes. S12: The catalyst particle sample is scanned using a laser three-dimensional scanning device to extract the surface roughness, number of edges and corners, sphericity, and surface porosity of the catalyst particle sample as morphological data of the catalyst particle sample; The catalyst particle sample was analyzed using a low-temperature nitrogen adsorption method to obtain the specific surface area, total pore volume, and average pore size of the catalyst particle sample, which were used as the pore structure data of the catalyst particle sample. The static packing density, tapped packing density, and bed porosity of the catalyst particle sample were measured and used as the packing density data of the catalyst particle sample. S13: Use the morphology data, pore structure data, and bulk density data of the catalyst particle sample as the original data of the sieve pore size associated with the catalyst particle sample. S14: Standardize the original data of the sieve aperture particle size, and weight the standardized morphology data to obtain the morphology index score of the sieve aperture particle size; weight the standardized pore structure data to obtain the pore structure index score of the sieve aperture particle size; weight the standardized bulk density data to obtain the bulk density index score of the sieve aperture particle size. The morphology score, pore structure score, and bulk density score of the sieve pore size are used as the characteristic levels of the sieve pore size. S15: The industrial bed reactor is divided into multiple bed space partitioning units using a three-dimensional mesh partitioning method. The unique number, axial start and end coordinates, radial start and end coordinates, circumferential start and end coordinates, center point coordinates, theoretical volume, and distance from the heat exchange tubes of each bed space partitioning unit are recorded as the spatial partitioning mapping result of the industrial bed reactor.

3. The artificial intelligence-driven catalytic industrial control method as described in claim 2, characterized in that, Step S1, which assigns baseline values ​​to the bed spatial partitioning units based on the target reaction conditions to construct the bed partitioning particle structure data, also includes: S16: Obtain the target reaction conditions in the catalytic process of the target reactant, wherein the target reaction conditions include hydrocarbon feed concentration, ammonia feed concentration, oxygen feed concentration, feed temperature, gas reactant flow rate and the maximum heat release per unit volume of catalyst bed, and summarize the hydrocarbon feed concentration, ammonia feed concentration and oxygen feed concentration as the total inlet flow rate; S17: Based on the distance between the bed space partitioning unit and the heat exchange tube, and the characteristic grade of the sieve particle size, combined with the total inlet flow rate and the maximum heat release per unit volume of catalyst bed, the operating condition reference value between the bed space partitioning unit and the catalyst particles associated with the sieve particle size is calculated, wherein the operating condition reference value includes the feed flow rate reference value and the heat release load reference value. S18: The spatial partitioning mapping result of the industrial bed reactor, the operating condition benchmark value between the bed spatial partitioning unit and the catalyst particles associated with the sieve aperture size, and the characteristic level of the sieve aperture size are used as the bed partitioning particle structure data.

4. The artificial intelligence-driven catalytic industrial control method as described in claim 1, characterized in that, In step S2, the transfer and reaction characteristics of the bed space partitioning unit are calculated in conjunction with the target reaction conditions, including: S21: Based on the particle structure data of the bed partition and the target reaction conditions, calculate the local gas flow obstruction intensity, local external diffusion transfer intensity and local internal diffusion transfer intensity of the bed space partition unit under the action of catalyst particles associated with the sieve aperture size, and use the local gas flow obstruction intensity, local external diffusion transfer intensity and local internal diffusion transfer intensity as the transfer characteristic quantities of the bed space partition unit under the action of catalyst particles associated with the sieve aperture size; S22: Calculate the local heat accumulation intensity, local ammonia dissipation rate, and unit intermediate residence intensity of the bed space partitioning unit under the action of catalyst particles associated with the sieve aperture size. Use the local heat accumulation intensity, local ammonia dissipation rate, and unit intermediate residence intensity as the reaction characteristic quantities of the bed space partitioning unit under the action of catalyst particles associated with the sieve aperture size.

5. The artificial intelligence-driven catalytic industrial control method as described in claim 4, characterized in that, Step S2 involves stitching together and reconstructing the bed state field and extracting the gradient distribution to obtain microscale reaction environment distribution data. This also includes: S23: Based on the aforementioned transfer and reaction characteristics, calculate the physical state quantities of the bed space partitioning unit under the action of catalyst particles associated with different sieve aperture sizes, wherein the physical state quantities include unit temperature value, unit ammonia concentration value, unit oxygen concentration value, and unit intermediate residence intensity. S24: Calculate the splicing weight between the bed space partitioning unit and the adjacent unit, and smoothly reconstruct the physical state quantity of the bed space partitioning unit to obtain the spatially continuous physical state quantity as the bed state field. S25: Based on the bed state field, extract the gradient difference index of the bed spatial partition unit as microscale reaction environment distribution data, wherein the gradient difference index includes temperature gradient index, ammonia concentration gradient index, oxygen concentration gradient index and intermediate gradient index.

6. The artificial intelligence-driven catalytic industrial control method as described in claim 1, characterized in that, In step S3, the reaction type mapping and risk coupling scoring of the bed spatial partitioning units are performed in conjunction with the catalytic reaction pathway, including: S31: Based on the microscale reaction environment distribution data, extract the gradient difference index of the bed spatial partition unit, and combine it with the catalytic reaction path to generate the reaction type of the bed spatial partition unit, wherein the reaction type includes the main reaction, amide side reaction and deep oxidation side reaction; S32: Based on the reaction type of the bed spatial partitioning unit, calculate the mean gradient difference index of the bed spatial partitioning unit under the action of catalyst particles associated with the U-group sieve pore size. Apply a risk coupling weighting to the mean gradient difference index of the bed spatial partitioning unit to obtain the risk coupling score of the bed spatial partitioning unit. This indicates the number of groups representing the sieve aperture size.

7. The artificial intelligence-driven catalytic industrial control method as described in claim 6, characterized in that, The formula for risk-coupled weighting of the gradient difference index of the bed space partitioning unit in step S32 is as follows: ； in, This represents the risk coupling score of the nth bed-level spatial partition unit. , indicating the total number of spatial partitioning units in the bed layer. This represents the mean gradient difference index of the nth bed-layer spatial partition unit. The values ​​represent the mean values ​​of the temperature gradient index, the ammonia concentration gradient index, the oxygen concentration gradient index, and the intermediate gradient index, respectively. This indicates the reaction type of the nth bed space partitioning unit.

8. The artificial intelligence-driven catalytic industrial control method as described in claim 1, characterized in that, In step S4, the bed partition particle structure data is called to match the characteristic level of the corresponding catalyst particles, and differentiated catalyst particle configuration is performed on the bed spatial partition unit, including: S41: Based on the risk coupling score of the bed space partitioning unit, the bed space partitioning unit is divided into high-risk unit, medium-risk unit and low-risk unit; S42: Based on the risk unit type and reaction type of the bed space partitioning unit, match catalyst particles with sieve pore size corresponding to the characteristic level as catalyst classification configuration data, and configure the catalyst particles matched by the bed space partitioning unit in the bed space partitioning unit.

9. The artificial intelligence-driven catalytic industrial control system as described in claim 1, characterized in that, The catalytic industrial control system includes a catalyst particle storage device and an industrial bed reactor. The catalyst particle storage device is used to store catalyst particles with multiple sets of sieve aperture sizes, and the industrial bed reactor is used to perform raw material conversion and target reactant generation, so as to realize the artificial intelligence-driven catalytic industrial control method as described in any one of claims 1-8.

Images