A catalytic cracking simulation system and method, system construction method and related device

Abstract

The application discloses a catalytic cracking simulation system and method, a system construction method and related devices. The construction method comprises: converting each chemical substance including raw oil and products in the catalytic cracking reaction into corresponding vector data, and storing the raw oil vector data in a pre-built database; constructing a digital twin model of the catalytic cracking reaction device; inputting the raw oil data and reaction condition data in the sample data into an initial catalytic cracking simulation system integrated by the database and the digital twin model, querying the raw oil data by using the database to obtain corresponding vector data, performing a catalytic cracking simulation reaction based on the vector data and the input reaction condition by using the digital twin model, and outputting a prediction result; verifying and optimizing the digital twin model in the initial catalytic cracking simulation system based on the prediction result and the true product result in the sample data, and obtaining a constructed catalytic cracking simulation system. The catalytic cracking simulation system can simulate the catalytic cracking reaction.

Description

Technical Field

[0001] This invention relates to the field of petrochemical technology, and in particular to a catalytic cracking simulation system and method, system construction method, and related apparatus. Background Technology

[0002] As a core piece of equipment in the petroleum refining process, the catalytic cracking unit's operating efficiency and product quality directly impact the economic benefits and environmental performance of refining enterprises. However, the catalytic cracking process involves a complex chemical reaction network, diverse feedstock compositions, and harsh reaction conditions, leading to numerous uncertainties during unit operation, such as coking in the settling tank, catalyst deactivation, equipment corrosion, and excessive emissions. Traditional operation methods relying on manual experience are no longer sufficient to meet the demands of modern refining industries for green, efficient, and intelligent production.

[0003] In recent years, with the rapid development of technologies such as big data, cloud computing, the Internet of Things, and artificial intelligence, digital transformation has become a key path to promote high-quality development in the petrochemical industry. Among them, digital twin technology, as an emerging technology, achieves holographic perception, intelligent analysis, and advanced prediction of the operating status of physical objects by constructing virtual mirrors, providing strong technical support for optimized control. Currently, digital twin technology for catalytic cracking units is still at the theoretical level. Therefore, how to apply digital twin technology to the simulation of catalytic cracking reactions is an urgent problem to be solved. Summary of the Invention

[0004] In view of the above problems, the present invention is proposed to provide a catalytic cracking simulation system and method, system construction method and related apparatus that overcome the above problems or at least partially solve the above problems.

[0005] In a first aspect, embodiments of the present invention provide a method for constructing a catalytic cracking simulation system, comprising:

[0006] The collected chemical substances from the catalytic cracking reaction are converted into their corresponding vector data. The chemical substances include feedstock oil and products. The vector data of the feedstock oil is stored in a pre-built database.

[0007] A digital twin model of catalytic cracking is constructed based on the vector data of each chemical substance, the physical data of the catalytic cracking reactor, and the reaction process of the catalytic cracking reaction.

[0008] The database and the digital twin model are integrated to obtain an initial catalytic cracking simulation system;

[0009] The feedstock oil data and reaction condition data from the sample data are input into the initial catalytic cracking simulation system. The input feedstock oil data is queried in the database to obtain the corresponding vector data. The digital twin model is used to simulate the catalytic cracking reaction based on the vector data and the input reaction conditions, and the predicted results are output. The digital twin model in the initial catalytic cracking simulation system is verified and optimized based on the predicted results and the corresponding real product results in the sample data, so as to obtain the constructed catalytic cracking simulation system.

[0010] In an optional embodiment, the collected chemical substances from the catalytic cracking reaction are converted into their corresponding vector data, including:

[0011] Based on the collected chemical substances in the catalytic cracking reaction, molecular models of each chemical substance in the catalytic cracking reaction were constructed.

[0012] Using selected molecular descriptors, the molecular models of each chemical substance are mapped to a vector space, and the corresponding vector data is determined based on the molecular descriptor values ​​of each chemical substance.

[0013] In an optional embodiment, based on the collected chemicals in the catalytic cracking reaction, a molecular model of each chemical substance in the catalytic cracking reaction is constructed; including:

[0014] Collect relevant information on the chemical substances involved in the catalytic cracking process, including their chemical composition and physical properties;

[0015] Based on the structural characteristics of the chemical composition and physical properties of each chemical substance, the chemical substances are classified, and one chemical substance is selected from each classification.

[0016] Based on the relevant information of the selected chemical substances, their three-dimensional structural information, chemical bond information and chemical properties are determined;

[0017] Based on the three-dimensional structural information, chemical bond information, and chemical properties of each selected chemical substance, a corresponding molecular model is constructed.

[0018] In an optional embodiment, the construction method further includes:

[0019] Based on feedstock and product data from catalytic cracking reaction experiments, the collected chemical substances were screened to remove feedstocks and products that did not belong to the catalytic cracking reaction.

[0020] In an optional embodiment, the construction method further includes: using machine learning algorithms to reduce the dimensionality of the vector data of each chemical substance.

[0021] In an optional embodiment, a digital twin model of catalytic cracking is constructed based on the vector data of each chemical substance, the physical data of the catalytic cracking reactor, and the reaction process of the catalytic cracking reaction, including:

[0022] A data twin model component is constructed based on the physical data of the catalytic cracking reactor. The reaction path of each reaction in the catalytic cracking reaction and its corresponding reactants and products are determined based on the reaction process of the catalytic cracking reaction. The reaction hierarchy of the digital twin model and the reactant interface corresponding to each reaction hierarchy are constructed based on the vector data of each reactant in the reaction process, its corresponding reaction path and the vector data of the corresponding products.

[0023] Develop an interaction interface between the digital twin model and an external system to achieve at least one of the following functions: receiving user input data, receiving real-time operating data of the catalytic cracking unit, and outputting prediction results.

[0024] In an optional embodiment, during the verification and optimization of the digital twin model, the sources of the feedstock data and reaction condition data in the sample data, and the corresponding real product results, include: the actual catalytic cracking reaction process;

[0025] And / or, input the preset feedstock oil data and reaction condition data into the pre-trained model, output the reaction result through the model, use the preset feedstock oil data and reaction condition data as the feedstock oil data and reaction condition data in the sample data, and use the reaction result output by the model as its corresponding real product result.

[0026] Based on the same inventive concept, embodiments of the present invention also provide a method for simulating a catalytic cracking reaction, comprising:

[0027] Obtain data on the feedstock and reaction conditions for the catalytic cracking reaction;

[0028] The feedstock data and reaction condition data to be simulated are input into a pre-constructed catalytic cracking simulation system. The database queries the input feedstock data to obtain the corresponding vector data. The digital twin model is used to perform catalytic cracking simulation based on the vector data and the input reaction condition data, and the prediction results are output.

[0029] The catalytic cracking reaction simulation system is constructed based on the above-described construction method.

[0030] In an optional embodiment, the database queries the input feedstock vector data to obtain its corresponding vector data, including:

[0031] The database uses a preset similarity calculation method to perform similarity calculations on the input raw oil data and the vector data stored in the database to obtain the vector data with the highest similarity to the raw oil vector data.

[0032] In an optional embodiment, the catalytic cracking reaction simulation method provided by the present invention further includes: analyzing the reaction path in the catalytic cracking reaction process and predicting the performance of the catalytic cracking unit based on the prediction results; the performance of the catalytic cracking unit includes: product distribution under different operating conditions, energy consumption level under different operating conditions, and pollutant emission under different operating conditions.

[0033] Based on the same inventive concept, embodiments of the present invention also provide a catalytic cracking simulation system, comprising: a database module and a digital twin model.

[0034] The database module is used to store vector data corresponding to each feedstock oil in the catalytic cracking reaction, and to query the corresponding vector data based on the input chemical substance data.

[0035] The digital twin model is used to simulate catalytic cracking reactions based on the vector data obtained from the database and the input reaction condition data, and outputs prediction results.

[0036] Based on the same inventive concept, embodiments of the present invention also provide an apparatus for constructing a catalytic cracking simulation system, comprising:

[0037] The first construction module is used to convert the collected chemical substances in the catalytic cracking reaction into their corresponding vector data, wherein the chemical substances include feedstock oil and products; and to store the vector data of the feedstock oil in a pre-built database.

[0038] The second construction module is used to construct a digital twin model of catalytic cracking based on the vector data of each chemical substance, the physical data of the catalytic cracking reactor, and the reaction process of the catalytic cracking reaction.

[0039] An integration module is used to integrate the database and the digital twin model to obtain an initial catalytic cracking simulation system;

[0040] The verification and optimization module is used to input feedstock oil data and reaction condition data from the sample data into the initial catalytic cracking simulation system, query the input feedstock oil data using a database to obtain corresponding vector data, use the digital twin model to perform catalytic cracking simulation reaction based on the vector data and the input reaction conditions, and output prediction results. Based on the prediction results and the corresponding real product results in the sample data, the digital twin model in the initial catalytic cracking simulation system is verified and optimized to obtain the constructed catalytic cracking simulation system.

[0041] Based on the same inventive concept, embodiments of the present invention also provide a catalytic cracking reaction simulation device, comprising: an acquisition module and a simulation module;

[0042] The acquisition module is used to acquire data on the feedstock and reaction conditions for the catalytic cracking reaction.

[0043] The simulation module is used to input the feedstock data and reaction condition data to be simulated into a pre-constructed catalytic cracking simulation system, query the input feedstock data using a database to obtain the corresponding vector data, use the digital twin model to perform catalytic cracking simulation reaction based on the vector data and the input reaction condition data, and output the prediction results.

[0044] The catalytic cracking reaction simulation system is constructed based on the above-described construction method.

[0045] Based on the same inventive concept, this embodiment of the invention also provides a computer storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method for constructing a catalytic cracking simulation system and the above-described method for simulating a catalytic cracking reaction.

[0046] Based on the same inventive concept, this invention also provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the above-described method for constructing a catalytic cracking simulation system and the above-described method for simulating a catalytic cracking reaction.

[0047] The beneficial effects of the above-described technical solutions provided in the embodiments of the present invention include at least the following:

[0048] This invention provides a method for constructing a catalytic cracking simulation system. The system comprises a database and a digital twin model of the catalytic cracking unit. During system construction, the chemical substances involved in the catalytic cracking reaction are converted into corresponding vector data, which are then stored in a pre-built database. In application, the corresponding vector data of the feedstock is queried through this vector database. Simultaneously, this method constructs a digital twin model of the catalytic cracking unit based on its physical data and the reaction flow of the catalytic cracking reaction. The database and the digital twin model are then integrated to obtain an initial catalytic cracking simulation system. Verification and optimization yield a fully constructed catalytic cracking simulation system. The catalytic cracking simulation system constructed using this method has a dedicated vector database for easy and rapid querying of vector data. In application, the catalytic cracking simulation system, through the constructed molecular-level refining model and the application of vector data in molecular refining, enables precise simulation and prediction of the operating status of the catalytic cracking unit.

[0049] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings.

[0050] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0051] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0052] Figure 1 This is a schematic flowchart of the method for constructing the catalytic cracking simulation system in Embodiment 1 of the present invention;

[0053] Figure 2 This is a schematic diagram illustrating the specific process of constructing the catalytic cracking simulation system in Embodiment 2 of the present invention;

[0054] Figure 3 This is a flowchart illustrating the molecular model construction steps in Embodiment 2 of the present invention;

[0055] Figure 4 This is a flowchart illustrating the vector data representation steps in Embodiment 2 of the present invention;

[0056] Figure 5 This is a flowchart illustrating the vector database design steps in Embodiment 2 of the present invention;

[0057] Figure 6 This is a schematic diagram of the process for constructing the digital twin model in Embodiment 2 of the present invention;

[0058] Figure 7 This is a flowchart illustrating the intelligent analysis and optimization steps in Embodiment 2 of the present invention;

[0059] Figure 8 This is a schematic diagram of the catalytic cracking simulation system in an embodiment of the present invention;

[0060] Figure 9 This is a schematic flowchart of the catalytic cracking reaction simulation method in an embodiment of the present invention;

[0061] Figure 10 This is a schematic diagram of the structure of the catalytic cracking simulation system construction device in an embodiment of the present invention;

[0062] Figure 11 This is a schematic diagram of the catalytic cracking simulation device in an embodiment of the present invention. Detailed Implementation

[0063] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0064] To address the challenge of applying digital twin technology to catalytic cracking reactions in existing technologies, this invention provides a catalytic cracking simulation system and method, a system construction method, and related apparatus. It should be noted that the input feedstock data in this embodiment is a set of vector data.

[0065] Example 1:

[0066] The method for constructing a catalytic cracking system provided in this embodiment of the invention has a flowchart as shown below. Figure 1 As shown, it includes the following steps:

[0067] Step S101: Convert the collected chemical substances in the catalytic cracking reaction into their corresponding vector data, wherein the chemical substances include: feedstock oil and products; and store the vector data of the feedstock oil in a pre-built database;

[0068] Step S102: Construct a digital twin model of catalytic cracking based on vector data of each chemical substance, physical data of the catalytic cracking reactor, and the reaction process of the catalytic cracking reaction;

[0069] Step S103: Integrate the database and the digital twin model to obtain an initial catalytic cracking simulation system;

[0070] Step S104: Input the feedstock data and reaction condition data from the sample data into the initial catalytic cracking simulation system. Use the database to query the input feedstock data to obtain the corresponding vector data. Use the digital twin model to perform a catalytic cracking simulation reaction based on the vector data and the input reaction conditions, and output the prediction results. Based on the prediction results and the corresponding real product results in the sample data, verify and optimize the digital twin model in the initial catalytic cracking simulation system to obtain the constructed catalytic cracking simulation system.

[0071] The feedstock oil data in the sample data may be a portion of the feedstock oil collected in the catalytic cracking reaction in step S101.

[0072] In one embodiment, the construction method provided by this invention, in step S101, converts each chemical substance collected in the catalytic cracking reaction into its corresponding vector data, including:

[0073] Based on the collected chemical substances in the catalytic cracking reaction, molecular models of each chemical substance in the catalytic cracking reaction were constructed.

[0074] Using selected molecular descriptors, the molecular models of each chemical substance are mapped to a vector space, and the corresponding vector data is determined based on the molecular descriptor values ​​of each chemical substance.

[0075] Furthermore, based on the collected chemical substances in the catalytic cracking reaction, molecular models of each chemical substance in the catalytic cracking reaction are constructed; including:

[0076] Collect relevant information on the chemical substances involved in the catalytic cracking process, including their chemical composition and physical properties;

[0077] Based on the structural characteristics of the chemical composition and physical properties of each chemical substance, the chemical substances are classified, and one chemical substance is selected from each classification.

[0078] Based on the relevant information of the selected chemical substances, their three-dimensional structural information, chemical bond information, and chemical properties are determined; wherein, the three-dimensional structural information includes: interatomic distance and bond angle; the chemical bond information includes: chemical bond type, bond length, and bond energy; and the chemical properties include: reactivity and reaction rate.

[0079] Based on the three-dimensional structural information, chemical bond information, and chemical properties of each selected chemical substance, a corresponding molecular model is constructed.

[0080] In catalytic cracking reactions, numerous chemical substances are involved. Building molecular models for all of these substances would be extremely labor-intensive. Classifying these substances not only reduces the workload of molecular model construction but also minimizes the possibility of not being able to retrieve the vector of the input feedstock during catalytic cracking simulations due to the absence of vector data for certain substances in the vector library. The classification of chemical substances can be selected according to actual needs, and this embodiment of the invention does not impose specific limitations on this. For example, chemical substances with the same molecular formula may have similar reaction properties; therefore, it is sufficient to select one of these substances with the same molecular formula and similar reaction properties for molecular model construction.

[0081] In one embodiment, the construction of a molecular model of a chemical substance further includes:

[0082] Based on feedstock and product data from catalytic cracking reaction simulations, the collected chemical substances were validated and screened to remove feedstocks and products that did not belong to the catalytic cracking reaction. Accordingly, molecular models were constructed for each of the screened chemical substances. Because the chemical substances collected in catalytic cracking reactions may originate from real catalytic cracking simulations, experimental measurements, computational simulations, or literature, the collected feedstock may be a mixture. Some substances in this mixture are not used in the catalytic cracking reaction, and some products are not catalytic cracking products. However, in catalytic cracking reaction simulations, the feedstock and products used are singular and accurate. Therefore, the collected data can be screened using real feedstock and product data used in the simulation, removing feedstock and product data that do not belong to the catalytic cracking reaction. This ensures the accuracy and reliability of the collected chemical substances and the subsequent molecular model construction, thereby guaranteeing the accuracy of the reactants and products determined in the catalytic cracking reaction process when constructing the digital twin model.

[0083] Furthermore, after obtaining the collected vector data corresponding to the transformations of each chemical substance, the process also includes: using machine learning algorithms to reduce the dimensionality of the vector data for each chemical substance; specifically, points in high-dimensional vector space may be difficult to intuitively understand and process. Machine learning algorithms, such as autoencoders or t-SNE, are used to map these high-dimensional vectors to a low-dimensional space (usually two-dimensional or three-dimensional) to facilitate visualization and further analysis. In the low-dimensional vector space, similar molecules should be spatially close to each other to reflect their similarity in chemical structure. This step ensures that the vector representation can capture the similarity relationships between molecules; at the same time, the low-dimensional vector representation makes molecular data easier to process and analyze. For example, molecular similarity searches can be performed quickly, or clustering algorithms can be applied in the low-dimensional space to discover groups of molecules with similar structural features.

[0084] Specifically, when using the t-SNE algorithm and autoencoder to reduce the dimensionality of vector data for each chemical substance, the processing principle is as follows:

[0085] 1. Calculate similarity in high-dimensional space: Based on the vector data corresponding to each chemical substance, map each vector data to data points in high-dimensional data, calculate the Euclidean distance between each pair of data points, and determine the similarity between each pair of data points based on the Euclidean distance;

[0086] 2. Calculate similarity in low-dimensional space: Randomly initialize the positions of data points in a low-dimensional space (usually two-dimensional or three-dimensional). Use gradient descent to optimize the positions of the molecules in the low-dimensional space to minimize the difference in similarity between data points in the high-dimensional and low-dimensional spaces. Specifically, the algorithm calculates a loss function that measures the difference between the conditional probabilities in the high-dimensional and low-dimensional spaces.

[0087] 3. Iterative position update: During gradient descent, the algorithm continuously updates the position of data points in the low-dimensional space to reduce the value of the loss function until it finds a position configuration that minimizes the loss function.

[0088] 4. Visualization results: After multiple iterations, high-dimensional data points are mapped to two-dimensional or three-dimensional space, and scatter plots or other visualization tools are used to show the distribution of these data points in low-dimensional space. Based on the distribution of these data points in low-dimensional space, the dimensionality reduction of each chemical substance results in vector data.

[0089] Furthermore, in step S102, a digital twin model of catalytic cracking is constructed based on the vector data of each chemical substance, the physical data of the catalytic cracking reactor, and the reaction process of the catalytic cracking reaction, including:

[0090] A data twin model component is constructed based on the physical data of the catalytic cracking reactor. The reaction path of each reaction in the catalytic cracking reaction and its corresponding reactants and products are determined based on the reaction process of the catalytic cracking reaction. The reaction hierarchy of the digital twin model and the reactant interface corresponding to each reaction hierarchy are constructed based on the vector data of each reactant in the reaction process, its corresponding reaction path and the vector data of the corresponding products. The reactant interface corresponding to each reaction hierarchy defines the reactants that undergo specific reactions.

[0091] Develop an interaction interface between the digital twin model and an external system to achieve at least one of the following functions: receiving user input data, receiving real-time operating data of the catalytic cracking unit, and outputting prediction results, so as to ensure that the digital twin model can receive input data, output prediction structures, and interact with the user; wherein, the external system may be: a data acquisition system, a user interface, a control system, etc.

[0092] In an optional embodiment, in step S104, when verifying and optimizing the digital twin model, the sources of the feedstock oil data and reaction condition data in the sample data, and the corresponding real product results, include: the actual catalytic cracking reaction process;

[0093] And / or, input the preset feedstock oil data and reaction condition data into the pre-trained model, output the reaction result through the model, use the preset feedstock oil data and reaction condition data as the feedstock oil data and reaction condition data in the sample data, and use the reaction result output by the model as its corresponding real product result.

[0094] In one embodiment, the construction method provided by this invention can construct a database through the following steps: determining the database architecture according to design requirements; wherein, the database architecture includes: the type of database model, the type of storage engine, and the index structure;

[0095] When building a database through a database architecture, preprocessing functions are integrated and similarity calculation methods are defined. The preprocessing functions include: data cleaning, deduplication, and standardization.

[0096] Test and optimize database performance to obtain the completed database.

[0097] The present invention provides a method for constructing a catalytic cracking simulation system. This system comprises a database and a digital twin model of the catalytic cracking unit. During system construction, the chemical substances involved in the catalytic cracking reaction are converted into corresponding vector data, which is then stored in a pre-built database. In application, the corresponding vector data of the feedstock is queried through this vector database. Simultaneously, this method constructs a digital twin model of the catalytic cracking based on the vector data, the physical data of the catalytic cracking unit, and the reaction flow of the catalytic cracking reaction. The database and the digital twin model are then integrated to obtain an initial catalytic cracking simulation system. Verification and optimization yield a completed catalytic cracking simulation system. The catalytic cracking simulation system constructed using this method has a dedicated vector database for easy and rapid querying of vector data. Furthermore, in application, the catalytic cracking simulation system, through the constructed molecular-level refining model and the application of vector data in molecular refining, can achieve precise simulation and prediction of the operating status of the catalytic cracking unit.

[0098] Example 2:

[0099] In one embodiment, the present invention provides a specific implementation process for the construction method of the above-mentioned catalytic cracking simulation system, the flowchart of which is shown below. Figure 2 As shown, it includes the following steps:

[0100] Step S201: Molecular model construction

[0101] This invention first constructs a molecular model library of catalytic cracking feedstocks and products based on theories of quantum mechanics, molecular mechanics, and statistical mechanics. This model library contains key parameters such as the three-dimensional structure, chemical bond information, and reactivity of various chemical substances involved in different catalytic cracking reactions, providing fundamental data for subsequent simulation calculations.

[0102] In one embodiment, refer to Figure 3 As shown, the detailed construction process of the molecular model is as follows:

[0103] 1) Determine the theoretical foundation: quantum mechanics, molecular mechanics, and statistical mechanics.

[0104] Quantum mechanics studies the behavior and properties of microscopic particles, providing fundamental theoretical support for molecular models.

[0105] Molecular mechanics: studies the mechanical properties of intermolecular interactions, providing crucial information for the construction of molecular models.

[0106] Statistical mechanics studies the relationship between the macroscopic properties and microscopic states of a large number of particle systems, providing a theoretical basis for the statistical analysis of molecular models.

[0107] 2) Collect chemical information on feedstocks and products of catalytic cracking. Specifically, collect information on the chemical composition and physical properties of feedstocks and products involved in the catalytic cracking process to provide basic data for the construction of molecular models.

[0108] 3) Constructing a molecular model library

[0109] a. Determine the classification and representativeness of the various chemical substances (generally various hydrocarbon molecules) in the catalytic cracking reaction.

[0110] Based on their chemical composition and structural characteristics, chemical substances are classified into different categories, and representative molecules are selected for model construction.

[0111] b. Construct the three-dimensional structure of each chemical substance.

[0112] Using quantum mechanics and molecular mechanics theories, the three-dimensional structure of chemical substances is calculated, including parameters such as interatomic distances and bond angles.

[0113] c. Determine chemical bond information

[0114] Analyzing the chemical bond types, bond lengths, and bond energies in various chemical substances provides key chemical bond parameters for molecular models.

[0115] d. Determine key parameters such as reactivity

[0116] Based on statistical mechanics and reaction kinetics, key parameters such as the reactivity and reaction rate of each chemical substance are calculated.

[0117] e. Based on the three-dimensional structural information, chemical bond information, and chemical properties of each chemical substance, construct the corresponding molecular model for each chemical substance.

[0118] 4) Verify the accuracy and reliability of the molecular model library.

[0119] The constructed molecular model library was screened using experimental data from catalytic cracking simulation reactions and existing theoretical models to ensure the accuracy and reliability of the models. The experimental data included feedstock data input into the existing theoretical models and product data output from the existing theoretical models.

[0120] Step S202: Vector Data Representation

[0121] To efficiently process molecular structure information, this invention employs a vector data representation method, transforming molecular structures into points in a high-dimensional vector space. Specifically, molecular descriptors (such as molecular fingerprints, the number of functional groups, and bond type distribution) are used as features, and machine learning algorithms (such as autoencoders and t-SNE) are employed to map the molecular structures into a low-dimensional vector space. This representation method preserves the similarity information between molecules while facilitating subsequent queries and calculations.

[0122] Specifically, refer to Figure 4 As shown, the detailed process of vector data representation is as follows:

[0123] 1) Determine the molecular descriptor

[0124] Molecular descriptors are a set of quantitative indicators used to characterize the structural features of molecules. These descriptors can be molecular fingerprints (a unique method of representing molecular structure), the number of functional groups (the number of specific chemical groups in a molecule), bond type distribution (the distribution of different types of chemical bonds in a molecule), etc. The selection of these descriptors should be able to comprehensively reflect the structural features of the molecule.

[0125] 2) Collect molecular structure information:

[0126] This step requires collecting detailed structural information about the molecules, including their chemical composition, three-dimensional spatial arrangement, and interatomic connections. This information forms the basis for vector data representation.

[0127] 3) Select a machine learning algorithm:

[0128] To effectively map the molecular structures of various chemical substances to a vector space, it is necessary to select appropriate machine learning algorithms. Autoencoders are unsupervised learning algorithms capable of learning effective encoded representations of data; t-SNE (t-distributed random neighborhood embedding) is an algorithm for high-dimensional data visualization, capable of mapping high-dimensional data to two- or three-dimensional space while preserving the similarity between data points.

[0129] 4) Mapping the molecular structure to a high-dimensional vector space:

[0130] Using selected molecular descriptors, the structural information of each molecule is transformed into a point in a high-dimensional vector space. This vector contains all the descriptor values ​​of the molecule, forming a mathematical representation of the molecule.

[0131] 5) Use machine learning algorithms to map high-dimensional vectors to low-dimensional vector spaces:

[0132] Points in high-dimensional vector spaces can be difficult to understand and process intuitively. Machine learning algorithms, such as autoencoders or t-SNE, can map these high-dimensional vectors to a low-dimensional space (usually two or three dimensions) to facilitate visualization and further analysis.

[0133] 6) Intermolecular similarity analysis:

[0134] In a low-dimensional vector space, similar molecules should be spatially close to each other to reflect their similarity in chemical structure. Transforming high-dimensional vectors of chemical substances into low-dimensional vectors through machine learning can preserve the similarity information between molecules, ensuring that the vector representation can capture the similarity relationships between molecules.

[0135] 7) Facilitates subsequent queries and calculations:

[0136] Low-dimensional vector representations make molecular data easier to process and analyze. For example, molecular similarity searches can be performed quickly, or clustering algorithms can be applied in low-dimensional space to discover groups of molecules with similar structural features.

[0137] In the step of representing vector data, 3) selecting a machine learning algorithm can be placed before or after 4) mapping the molecular structure to a high-dimensional vector space. This embodiment of the invention does not specify this in detail; for example, refer to... Figure 4 The diagram shown is a flowchart of the process before 3) selecting a machine learning algorithm and placing it in 4) mapping the molecular structure to a high-dimensional vector space.

[0138] Step S203: Vector Database Design

[0139] This invention provides a dedicated vector database system for storing and managing vector data in catalytic cracking processes. This database system supports efficient data indexing and querying, enabling rapid retrieval of sets of molecules similar to a given molecular vector and calculation of their similarity. Furthermore, the database integrates preprocessing functions such as data cleaning, deduplication, and standardization to ensure data accuracy and consistency.

[0140] Reference Figure 5 As shown, the specific process for constructing a vector database can be described as follows:

[0141] 1) Design the database architecture:

[0142] a. Determine the data model: Based on the characteristics of vector data, select an appropriate data model. For example, if the data has a fixed pattern and structure, a relational database model can be chosen; if the data structure is flexible and varied, a non-relational database model can be chosen.

[0143] b. Choose a storage engine: Select a suitable storage engine based on your data model and performance requirements. For example, in-memory databases offer fast read and write capabilities, suitable for scenarios requiring high-speed access; disk databases offer greater storage capacity and cost-effectiveness.

[0144] c. Design the index structure: To improve query efficiency, design a suitable index structure. For example, inverted indexes are suitable for quickly retrieving data related to specific keywords; spatial indexes are suitable for handling high-dimensional spatial data, such as molecular vectors.

[0145] 2) Integrated preprocessing function:

[0146] a. Data cleaning: Removing noise and outliers from data through preprocessing steps to improve data quality.

[0147] b. Deduplication: Identify and delete duplicate vector data to reduce data redundancy and improve storage efficiency.

[0148] c. Standardization: Standardize the vector data to eliminate the influence of different dimensions and ensure data consistency.

[0149] 3) Implement data indexing and query operations:

[0150] a. Support efficient data indexing: Implement an efficient data indexing mechanism to support fast data retrieval. This may include creating indexes, maintaining indexes, and optimizing index queries.

[0151] b. Implement a fast query algorithm: Based on the index structure, implement a fast query algorithm to retrieve a set of molecules similar to a given molecule vector. The query algorithm used may include nearest neighbor search algorithms, range search algorithms, etc.

[0152] 4) Storing and managing vector data:

[0153] a. Collect vector data: Collect vector data of each chemical substance from the catalytic cracking process. The data of each chemical substance may come from experimental measurements, calculation simulations or literature. Specifically, the vector data can be obtained through the methods of steps S201 and S202 above.

[0154] b. Store the data in a database: Store the collected vector data in the designed database to ensure that the data organization and storage methods can support efficient querying and analysis.

[0155] 5) Integrates similarity calculation function between molecular vectors:

[0156] a. Define the similarity calculation method: Select an appropriate similarity calculation method based on business needs. For example, cosine similarity is suitable for measuring the angular difference between vectors, while Euclidean distance is suitable for measuring the straight-line distance between vectors.

[0157] b. Implement similarity calculation function: Implement similarity calculation function in the database system to calculate the similarity between molecular vectors, supporting subsequent data analysis and pattern recognition.

[0158] 6) Test and optimize database performance:

[0159] a. Conduct performance testing: Perform performance testing on the database system to evaluate its query speed, index efficiency, data throughput, and other performance indicators.

[0160] b. Optimize based on test results: Based on the performance test results, optimize the database system to improve its performance and stability. This may include adjusting the index structure, optimizing query algorithms, and increasing hardware resources.

[0161] Specifically, when constructing the database, existing technologies can be combined to construct it according to actual needs, and the embodiments of the present invention do not impose specific limitations on this.

[0162] Step S204: Digital Twin Model Construction

[0163] Based on the aforementioned molecular model and vector database, this invention constructs a digital twin model of a catalytic cracking unit. This model predicts product distribution, energy consumption levels, and pollutant emissions under different operating conditions by simulating intermolecular interactions and reaction pathways. Furthermore, the model integrates real-time data interaction capabilities, enabling it to dynamically receive actual operating data from the unit and update model parameters accordingly, achieving adaptive optimization of the model.

[0164] Reference Figure 6 As shown, the specific process of constructing a digital twin model and its related applications can be illustrated by the following example:

[0165] 1) Prepare molecular model and vector database data:

[0166] a. Collect molecular model data and prepare vector database data to provide basic data for digital twin models. The molecular model library includes key parameters such as molecular structure, chemical bond information, and reactivity, while the vector database contains vector data related to the molecular models. This vector data may include features such as molecular fingerprints, the number of functional groups, and bond type distribution.

[0167] Furthermore, the molecular model and vector data are specifically constructed in steps S201 and 202 of this embodiment.

[0168] 2) Constructing a digital twin model framework:

[0169] a. Define the model structure: Design the basic structure of the digital twin model, including the model hierarchy, components, reactant interfaces corresponding to the reaction levels, etc., to ensure that the model can fully reflect the characteristics of the catalytic cracking unit.

[0170] b. Design the model interface: Develop the interaction interface between the model and external systems (such as data acquisition systems, user interfaces, control systems, etc.) to ensure that the model can receive input data, output prediction results, and interact with users.

[0171] 3) Simulate intermolecular interactions and reaction pathways:

[0172] a. Simulation using vector data: Using vector data, computational chemistry or physics methods are used to simulate intermolecular interactions, such as molecular collisions, the formation and breaking of chemical bonds, etc.

[0173] b. Determine the reaction pathway: Based on the simulation results, analyze the possible reaction pathways in the catalytic cracking process, including the main reaction and side reactions, and their impact on product distribution.

[0174] 4) Prediction device performance:

[0175] a. Predict product distribution: Based on the reaction pathway and molecular model, predict the types and proportions of products from the catalytic cracking unit under different operating conditions.

[0176] b. Predict energy consumption levels: Assess energy consumption levels under different operating conditions, including energy consumption of processes such as reaction heat, heat transfer, and mass transfer.

[0177] c. Predict pollutant emissions: Analyze pollutant emissions under different operating conditions, such as SOx, NOx, and particulate matter, to provide a basis for environmental protection measures.

[0178] 5) Integrates real-time data interaction functionality:

[0179] a. Design of data receiving interface: When constructing the data twin model, an interface was developed to receive actual operating data of the catalytic cracking unit, such as parameters like temperature, pressure, and flow rate.

[0180] b. Achieve dynamic data updates: Based on the received real-time data, dynamically update the parameters of the digital twin model to reflect the actual operating status of the catalytic cracking unit.

[0181] 6) Achieve adaptive optimization of the model:

[0182] a. Update model parameters based on real-time data: This enables the model to automatically adjust parameters based on real-time data to adapt to changes in the operating status of the catalytic cracking unit.

[0183] b. Optimize the model to adapt to different operating conditions: Based on the model's prediction results and actual operating data, optimize the model's structure and parameters to improve the model's prediction accuracy and adaptability.

[0184] 7) Verify and test the digital twin model:

[0185] a. Conduct model verification tests: Verify the accuracy, reliability, and stability of the digital twin model by comparing it with the actual device operation data.

[0186] b. Adjust and optimize based on test results: Based on the verification test results, make necessary adjustments and optimizations to the digital twin model to improve model performance and ensure that the model can meet the needs of practical applications.

[0187] In one embodiment, refer to Figure 7 As shown in the figure, the specific construction process of the model used to obtain sample data, and its specific application in verifying and optimizing the digital twin model, are described in the following embodiments of the present invention:

[0188] This invention utilizes big data and artificial intelligence technologies to intelligently analyze the operational status of digital twin models. By comparing the predicted results of the digital twin model with actual operational data, it identifies anomalies and potential risks in the operation of the digital twin model within the catalytic cracking simulation system. Based on this, it proposes targeted optimization suggestions and operational plans to guide the optimization of the digital twin model.

[0189] 1) Collect operating data of the catalytic cracking unit:

[0190] a. Real-time data acquisition: Key operating parameters such as temperature, pressure, and flow rate are collected in real time through sensors and data acquisition systems installed on the catalytic cracking unit.

[0191] b. Historical data organization: Organize and analyze the historical operating data of the device, including operation logs, maintenance records, fault reports, etc., to provide a reference for subsequent data analysis and model training.

[0192] 2) Applying big data technology for data processing:

[0193] a. Data cleaning and preprocessing: Clean the collected data, remove outliers, fill in missing values, and perform standardization to improve data quality.

[0194] b. Data storage and management: Store the cleaned and preprocessed data in the relevant database and manage it effectively to facilitate subsequent analysis and querying.

[0195] 3) Utilize artificial intelligence technology for intelligent analysis:

[0196] a. Model training and validation: Select appropriate machine learning algorithms, such as neural networks and support vector machines, to model the data, and train and validate the model through methods such as cross-validation.

[0197] b. Anomaly Detection and Risk Assessment: The trained model is used to monitor the operation status of the digital twin model in the catalytic cracking simulation system in real time, identify anomalies and potential risks in the constructed digital twin model, and provide a basis for subsequent optimization.

[0198] 4) Compare the predictions of the digital twin model with the actual operating data:

[0199] a. Data Comparison and Analysis: Based on the operating data of the catalytic cracking unit, the prediction results of the digital twin model are compared with the actual operating data obtained through the model to analyze the consistency and differences between the two.

[0200] b. Identify differences and deviations: Through comparative analysis, identify the differences and deviations between the predictions of the digital twin model and the actual operation, providing specific directions for the optimization of the digital twin model.

[0201] Example 3:

[0202] This invention uses a 2.6 Mt / a catalytic cracking unit as an example to illustrate the implementation process and effects of the above-mentioned simulation system construction as follows:

[0203] 1. Preliminary preparations

[0204] Data Acquisition: First, real-time operating data of the catalytic cracking unit is collected through the existing DCS (Distributed Control System) and sensor network, including key parameters such as feedstock composition, operating temperature, pressure, and catalyst activity.

[0205] Molecular model library construction: Based on the laboratory-tested feedstock oil samples, a molecular model library of feedstock oils and expected products is constructed using chemical analysis software (such as Gaussian, Materials Studio, etc.), and the corresponding molecular descriptors are calculated.

[0206] 2. Establishment of Vector Database

[0207] Data preprocessing: Convert molecular descriptors into vector representations and perform standardization to ensure that data from different sources have the same scale.

[0208] Database Design: A dedicated vector database system was designed and implemented using efficient vector indexing techniques (such as FAISS and Annoy). This system supports fast querying and similarity calculation, and can quickly find sets of molecules similar to a given molecular vector.

[0209] 3. Digital Twin Model Construction

[0210] Model integration: The pre-built digital twin model and vector database are integrated to obtain a catalytic cracking simulation system, enabling molecular-level simulation calculations.

[0211] Real-time data interaction: Develop a real-time data interface for the digital twin model, connect the DCS system of the catalytic cracking unit with the digital twin model, and realize real-time data transmission and synchronous updates.

[0212] 4. Intelligent analysis and optimization of catalytic cracking units.

[0213] Operational status monitoring: The catalytic cracking simulation system is used to monitor the operational status of the catalytic cracking unit in real time. By comparing the simulation results with actual data, potential abnormal operating conditions can be identified. For example, by comparing the actual data and predicted results of intermediate products in the reaction, it can be determined whether a certain sub-unit in the catalytic cracking unit is malfunctioning.

[0214] Early warning and decision support: When abnormal operating conditions are detected, relevant operators can analyze possible causes of failure and optimization suggestions based on the abnormal conditions, and adjust the operating parameters of the catalytic cracking unit based on this information to avoid unplanned shutdowns.

[0215] Energy saving and emission reduction optimization: Based on the prediction results of the simulation system, parameters such as the amount of alkali added, ozone injection rate, stripping steam and atomizing steam flow rate during the catalytic cracking reaction of the 2.6 Mt / a catalytic cracking unit can be optimized to achieve energy saving and emission reduction targets.

[0216] Based on the same inventive concept, embodiments of the present invention also provide a catalytic cracking simulation system, referring to... Figure 8 As shown, it includes: a database module 11 and a digital twin model 12.

[0217] The database module 11 is used to store vector data corresponding to each chemical substance in the catalytic cracking reaction, and to query the corresponding vector data based on the input chemical substance; wherein the chemical substances include: feedstock oil and products;

[0218] The digital twin model 12 is used to simulate catalytic cracking reaction based on the vector data obtained from the database and the input reaction condition data, and output the prediction results.

[0219] In one embodiment, the catalytic cracking simulation system of this invention is constructed based on the above-described catalytic cracking simulation system construction method. The functions implemented by each module have been described in the above embodiments, and will not be elaborated in detail in this embodiment.

[0220] Based on the same inventive concept, embodiments of the present invention also provide a method for simulating catalytic cracking reactions, referring to... Figure 9 As shown, it includes the following steps:

[0221] Step S301: Obtain the feedstock data and reaction condition data for the catalytic cracking reaction;

[0222] Step S302: Input the feedstock data and reaction condition data to be simulated into the pre-constructed catalytic cracking simulation system. The database queries the input feedstock data to obtain the corresponding vector data. A digital twin model is then used to simulate the catalytic cracking reaction based on the vector data and the input reaction condition data, and the predicted results are output. The predicted results include: each product in the catalytic cracking reaction; the products include: intermediate products and final products.

[0223] The catalytic cracking reaction simulation system is constructed based on the above-described construction method.

[0224] Specifically, the data on feedstock and reaction conditions for the catalytic cracking reaction to be simulated can be data input by the user as needed, or real-time operating data from the actual catalytic cracking unit received through the interactive interface of the digital twin model.

[0225] In one embodiment, the database queries the input feedstock vector data to obtain its corresponding vector data, including:

[0226] The database uses a preset similarity calculation method to perform similarity calculations on the input raw oil vector data and the vector data stored in the database to obtain the vector data with the highest similarity to the raw oil vector data.

[0227] In one embodiment, the above-mentioned catalytic cracking reaction simulation method further includes: analyzing the reaction path in the catalytic cracking reaction process and predicting the performance of the catalytic cracking unit based on the prediction results; the performance of the catalytic cracking unit includes: product distribution under different operating conditions, energy consumption level under different operating conditions, and pollutant emission under different operating conditions.

[0228] Specifically, when using a catalytic cracking simulation system to simulate catalytic cracking reactions, the digital twin model uses vector data of the feedstock oil to simulate intermolecular interactions, such as molecular collisions, the formation and breaking of chemical bonds, through computational chemistry or physical methods, and outputs simulation results.

[0229] Based on the simulation results, possible reaction pathways in the catalytic cracking process can be analyzed, including major and side reactions, and their impact on product distribution.

[0230] Simultaneously, using this simulation system, the following predictions can be made based on user-input data of feedstock oil and reaction conditions under different operating conditions:

[0231] a. Predict product distribution: Based on the reaction pathway and molecular model, predict the types and proportions of products from the catalytic cracking unit under different operating conditions.

[0232] b. Predict energy consumption levels: Assess energy consumption levels under different operating conditions, including energy consumption of processes such as reaction heat, heat transfer, and mass transfer.

[0233] c. Predict pollutant emissions: Analyze pollutant emissions under different operating conditions, such as SOx, NOx, and particulate matter, to provide a basis for environmental protection measures.

[0234] In one embodiment, the specific construction process of the catalytic cracking simulation system can refer to the construction process of the digital twin model in the above embodiment, and will not be specifically described in this embodiment.

[0235] Based on the same inventive concept, embodiments of the present invention also provide an apparatus for constructing a catalytic cracking simulation system, referring to... Figure 10 As shown, it includes:

[0236] The first construction module 21 is used to convert the collected chemical substances in the catalytic cracking reaction into their corresponding vector data, and store the vector data in a pre-built database; the chemical substances include: feedstock oil and products;

[0237] The second construction module 22 is used to construct a digital twin model of catalytic cracking based on the vector data, the physical data of the catalytic cracking reactor, and the reaction process of the catalytic cracking reaction.

[0238] Integration module 23 is used to integrate the database and the digital twin model to obtain an initial catalytic cracking simulation system;

[0239] The verification and optimization module 24 is used to input the feedstock oil data and reaction condition data from the sample data into the initial catalytic cracking simulation system, query the input feedstock oil data using the database to obtain the corresponding vector data, use the digital twin model to perform catalytic cracking simulation reaction based on the vector data and the input reaction conditions, and output the prediction results. Based on the prediction results and the corresponding real product results in the sample data, the digital twin model in the initial catalytic cracking simulation system is verified and optimized to obtain the constructed catalytic cracking simulation system.

[0240] It should be noted that the specific methods by which each module performs its operation in the construction apparatus of the catalytic cracking simulation system in the above embodiments have been described in detail in the embodiments of the relevant method, and will not be elaborated here.

[0241] Based on the same inventive concept, embodiments of the present invention also provide a simulation apparatus for a catalytic cracking reaction, referring to... Figure 11As shown, it includes:

[0242] The acquisition module 31 is used to acquire the feedstock oil data and reaction condition data of the catalytic cracking reaction to be simulated;

[0243] The simulation module 32 is used to input the simulated feedstock data and reaction condition data into a pre-constructed catalytic cracking simulation system, query the input feedstock data using a database to obtain the corresponding vector data, use the digital twin model to perform catalytic cracking simulation reaction based on the vector data and the input reaction condition data, and output the prediction results.

[0244] The catalytic cracking reaction simulation system is constructed based on the above-described construction method.

[0245] It should be noted that the specific methods by which each module performs its operation in the simulation apparatus for the catalytic cracking reaction in the above embodiments have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0246] Based on the same inventive concept, this embodiment of the invention also provides a computer storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method for constructing a catalytic cracking simulation system and the above-described method for simulating a catalytic cracking reaction.

[0247] Based on the same inventive concept, this invention also provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the above-described method for constructing a catalytic cracking simulation system and the above-described method for simulating a catalytic cracking reaction.

[0248] The application of the catalytic cracking simulation system and method, system construction method, and related apparatus provided in this invention has significantly improved the operating efficiency and safety of catalytic cracking units. This technology utilizes advanced molecular simulation, big data analysis, and artificial intelligence algorithms, combined with actual operating data from the catalytic cracking unit, to construct a high-precision molecular refining digital twin model. The application of digital twin technology enables comprehensive monitoring and intelligent management of the unit's operating status, enhancing risk perception, comprehensive analysis, and intelligent decision-making capabilities. The establishment and application of the vector database optimizes data storage and retrieval efficiency, improves the accuracy of similarity calculations, and provides strong support for the prediction of the digital twin model.

[0249] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0250] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0251] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0252] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0253] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for constructing a catalytic cracking simulation system, characterized in that, include: The collected chemical substances from the catalytic cracking reaction are converted into their corresponding vector data. The chemical substances include feedstock oil and products. The vector data of the feedstock oil is stored in a pre-built database. A digital twin model of catalytic cracking is constructed based on the vector data of each chemical substance, the physical data of the catalytic cracking reactor, and the reaction process of the catalytic cracking reaction. The database and the digital twin model are integrated to obtain an initial catalytic cracking simulation system; The feedstock oil data and reaction condition data from the sample data are input into the initial catalytic cracking simulation system. The input feedstock oil data is queried in the database to obtain the corresponding vector data. The digital twin model is used to simulate the catalytic cracking reaction based on the vector data and the input reaction conditions, and the predicted results are output. The digital twin model in the initial catalytic cracking simulation system is verified and optimized based on the predicted results and the corresponding real product results in the sample data, so as to obtain the constructed catalytic cracking simulation system.

2. The construction method as described in claim 1, characterized in that, The collected chemical substances from the catalytic cracking reaction were converted into their corresponding vector data, including: Based on the collected chemical substances in the catalytic cracking reaction, molecular models of each chemical substance in the catalytic cracking reaction were constructed. Using selected molecular descriptors, the molecular models of each chemical substance are mapped to a vector space, and the corresponding vector data is determined based on the molecular descriptor values ​​of each chemical substance.

3. The construction method as described in claim 2, characterized in that, Based on the collected chemical substances in the catalytic cracking reaction, molecular models of each chemical substance in the catalytic cracking reaction were constructed; including: Collect relevant information on the chemical substances involved in the catalytic cracking process, including their chemical composition and physical properties; Based on the structural characteristics of the chemical composition and physical properties of each chemical substance, the chemical substances are classified, and one chemical substance is selected from each classification. Based on the relevant information of the selected chemical substances, their three-dimensional structural information, chemical bond information and chemical properties are determined; Based on the three-dimensional structural information, chemical bond information, and chemical properties of each selected chemical substance, a corresponding molecular model is constructed.

4. The construction method as described in claim 1, characterized in that, Also includes: Based on feedstock and product data from catalytic cracking reaction experiments, the collected chemical substances were screened to remove feedstocks and products that did not belong to the catalytic cracking reaction.

5. The construction method as described in claim 2, characterized in that, Also includes: Machine learning algorithms are used to reduce the dimensionality of the vector data of each chemical substance.

6. The construction method as described in claim 1, characterized in that, A digital twin model of catalytic cracking is constructed based on the vector data of each chemical substance, the physical data of the catalytic cracking reactor, and the reaction process of the catalytic cracking reaction, including: A data twin model component is constructed based on the physical data of the catalytic cracking reactor. The reaction path of each reaction in the catalytic cracking reaction and its corresponding reactants and products are determined based on the reaction process of the catalytic cracking reaction. The reaction hierarchy of the digital twin model and the reactant interface corresponding to each reaction hierarchy are constructed based on the vector data of each reactant in the reaction process, its corresponding reaction path and the vector data of the corresponding products. Develop an interaction interface between the digital twin model and an external system to achieve at least one of the following functions: receiving user input data, receiving real-time operating data of the catalytic cracking unit, and outputting prediction results.

7. The construction method according to any one of claims 1-6, characterized in that, When validating and optimizing the digital twin model, the sources of the feedstock oil data and reaction condition data in the sample data, as well as the corresponding real product results, include: the actual catalytic cracking reaction process; And / or, input the preset feedstock oil data and reaction condition data into the pre-trained model, output the reaction result through the model, use the preset feedstock oil data and reaction condition data as the feedstock oil data and reaction condition data in the sample data, and use the reaction result output by the model as its corresponding real product result.

8. A method for simulating a catalytic cracking reaction, characterized in that, include: Obtain data on the feedstock and reaction conditions for the catalytic cracking reaction; The feedstock data and reaction condition data to be simulated are input into a pre-constructed catalytic cracking simulation system. The database queries the input feedstock data to obtain the corresponding vector data. The digital twin model is used to perform catalytic cracking simulation based on the vector data and the input reaction condition data, and the prediction results are output. The catalytic cracking reaction simulation system is constructed based on the construction method described in any one of claims 1-7.

9. The simulation method as described in claim 8, wherein the database queries the input feedstock vector data to obtain the corresponding vector data, including: The database uses a preset similarity calculation method to perform similarity calculations on the input raw oil data and the vector data stored in the database to obtain the vector data with the highest similarity to the raw oil vector data.

10. The simulation method as described in claim 8 or 9, characterized in that, It also includes: based on the prediction results, analyzing the reaction pathway in the catalytic cracking process and predicting the performance of the catalytic cracking unit; the performance of the catalytic cracking unit includes: product distribution under different operating conditions, energy consumption level under different operating conditions, and pollutant emissions under different operating conditions.

11. A catalytic cracking simulation system, characterized in that, include: Database module and digital twin model: The database module is used to store vector data corresponding to each feedstock oil in the catalytic cracking reaction, and to query the corresponding vector data based on the input chemical substance data. The digital twin model is used to simulate catalytic cracking reactions based on the vector data obtained from the database and the input reaction condition data, and outputs prediction results.

12. An apparatus for constructing a catalytic cracking simulation system, characterized in that, include: The first construction module is used to convert the collected chemical substances in the catalytic cracking reaction into their corresponding vector data, wherein the chemical substances include feedstock oil and products; and to store the vector data of the feedstock oil in a pre-built database. The second construction module is used to construct a digital twin model of catalytic cracking based on the vector data of each chemical substance, the physical data of the catalytic cracking reactor, and the reaction process of the catalytic cracking reaction. An integration module is used to integrate the database and the digital twin model to obtain an initial catalytic cracking simulation system; The verification and optimization module is used to input feedstock oil data and reaction condition data from the sample data into the initial catalytic cracking simulation system, query the input feedstock oil data using a database to obtain corresponding vector data, use the digital twin model to perform catalytic cracking simulation reaction based on the vector data and the input reaction conditions, and output prediction results. Based on the prediction results and the corresponding real product results in the sample data, the digital twin model in the initial catalytic cracking simulation system is verified and optimized to obtain the constructed catalytic cracking simulation system.

13. A catalytic cracking reaction simulation device, characterized in that, include: Acquisition module and simulation module; The acquisition module is used to acquire data on the feedstock and reaction conditions for the catalytic cracking reaction. The simulation module is used to input the feedstock data and reaction condition data to be simulated into a pre-constructed catalytic cracking simulation system, query the input feedstock data using a database to obtain the corresponding vector data, use the digital twin model to perform catalytic cracking simulation reaction based on the vector data and the input reaction condition data, and output the prediction results. The catalytic cracking reaction simulation system is constructed based on the construction method described in any one of claims 1-7.

14. A computer storage medium, characterized in that, The computer storage medium stores computer-executable instructions, which, when executed by a processor, implement the method for constructing the catalytic cracking simulation system according to any one of claims 1-7 and the method for simulating the catalytic cracking reaction according to any one of claims 8-10.

15. A computer device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method for constructing the catalytic cracking simulation system according to any one of claims 1-7 and the method for simulating the catalytic cracking reaction according to any one of claims 8-10.

Images