Refining simulation system and method for simulating molecular level reactant streams of a refining process

Abstract

The application discloses a kind of oil refining simulation system and the method for oil refining process molecular level reactant flow simulation.The system includes: molecular composition analysis model, secondary processing device model and blending optimization model;Molecular composition analysis model is used to predict the molecular composition and content of crude oil sample based on the physical data and characteristic data of crude oil sample and output using the trained machine learning algorithm model;Secondary processing device model is used to predict the molecular composition and physical property parameters of reaction product according to the molecular composition and content of crude oil sample and output;Blending optimization model is used to obtain the optimal value of blending ratio according to the molecular composition, physical property parameters and target physical property parameters of product, and simulate blending operation using the optimal value of blending ratio.The application models the whole process of oil refining process from the molecular level, realizes the accurate simulation of oil refining process, improves the adaptability and efficiency of oil refining process, and provides an analysis tool for the optimization of oil refining process.

Description

Technical Field

[0001] This invention relates to the field of computer simulation technology, and in particular to a refining simulation system and a method for simulating molecular-level reactive material flow in the refining process. Background Technology

[0002] In the petrochemical industry, refining process simulation software and digital models are in a stage of rapid development and play a key role in promoting industry progress.

[0003] The types of simulation software and digital models are becoming increasingly diverse. For example, process flow simulation software can accurately calculate and optimize the entire refining process; multiphase flow simulation software can effectively simulate complex fluid flows, providing technical support for equipment design and process optimization; and various types of digital models can achieve real-time monitoring of production, fault diagnosis, and optimization decision-making, improving production safety and economic efficiency. However, some problems remain to be solved. Traditional refining simulation software is mainly based on structure-oriented lumped (SOL) models, which have adaptability issues when dealing with new feedstocks and catalysts. Furthermore, although progress has been made in the development of molecular-level process models for individual reaction systems, there is a lack of molecular-level modeling for the entire process, making it impossible to achieve full-process molecular-level simulation of processing units, and a complete and unified system has not yet been formed. Summary of the Invention

[0004] In view of the above problems, the present invention is proposed to provide a refining simulation system and a method for simulating molecular-level reactive mass flow in the refining process to overcome or at least partially solve the above problems.

[0005] In a first aspect, embodiments of the present invention provide a refining simulation system, comprising: a molecular composition analysis model, a secondary processing device model, and a blending optimization model; wherein:

[0006] The molecular composition analysis model is used to receive the physical property data and chemical analysis characteristic data of the input crude oil sample, and use the trained machine learning algorithm model to predict the molecular composition and content of the crude oil sample and output it to the secondary processing device model.

[0007] The secondary processing device model is used to simulate the reaction process of the crude oil sample in the secondary processing device based on the molecular composition and content of the crude oil sample output by the molecular composition analysis model, predict the molecular composition and physical property parameters of the reaction products, and output them to the harmonization optimization model.

[0008] The blending optimization model is used to calculate and iteratively adjust the blending ratio of the blended product based on the molecular composition and physical property parameters output by the secondary processing device model and the preset target physical property parameters of the blended product, until it meets the preset target physical property parameters, obtain the optimal value of the blending ratio, and use the optimal value of the blending ratio to simulate the blending operation.

[0009] In one embodiment, the harmonic optimization model is pre-built and optimized in the following manner:

[0010] Based on the molecular composition and physical property parameters output by the secondary processing device model, the initial value of the blending ratio is calculated using an optimized mathematical algorithm.

[0011] Based on the initial blending ratio, the physical properties of the blended product are predicted. The blending result is evaluated by comparing the predicted physical properties of the blended product with the preset target physical properties. If the difference is greater than the preset difference threshold, the blending ratio is adjusted, and the physical properties of the blended product are predicted again based on the adjusted blending ratio. The blending result is evaluated again based on the difference between the predicted physical properties of the blended product and the target physical properties, until the difference is lower than the preset difference threshold, thus obtaining the optimal blending ratio.

[0012] In one embodiment, the harmonization optimization model is specifically constructed by employing a mathematical optimization algorithm. Based on the molecular composition and physical property parameters of the reaction products output by the secondary processing device model, the model aims to minimize the weighted square difference between the target physical property parameters and the physical property parameters output by the secondary processing device model. Combining the physical property parameters and content constraints of each component, the model numerically solves for the optimal solution to obtain the initial value of the harmonization ratio.

[0013] In one embodiment, the secondary processing device model is pre-built and trained in the following manner:

[0014] Construct a molecular composition matrix for the raw materials; the molecular composition matrix is ​​a matrix describing the composition of each molecule in the raw materials, with each row representing a type of molecule and each column representing the number of atoms of different elements in the molecule;

[0015] Based on the molecular composition matrix of the raw materials, and combined with the reaction rules and kinetic equations of catalytic cracking, a mathematical model is constructed to describe the relationship between the reaction rate and the concentrations of reactants and products under different reaction rules in the secondary processing unit.

[0016] By comparing the actual products corresponding to the raw materials with the raw material molecular composition matrix, the kinetic parameters in the mathematical model are continuously adjusted to minimize the difference between the predicted products and the actual products until the preset accuracy requirements are met.

[0017] In one embodiment, the molecular composition analytical model is pre-constructed and trained in the following manner:

[0018] The molecular structure and content of each component in the crude oil sample were determined to obtain data on the molecular composition of the crude oil sample.

[0019] Characteristic data of the molecular composition of crude oil samples were obtained through chemical analysis;

[0020] The data on the molecular composition, molecular composition features, and physical properties of the crude oil sample are used to form a dataset for model training.

[0021] Using the dataset, a pre-architected machine learning algorithm model is trained. During the training process, the parameters of the machine learning algorithm model are adjusted so that the machine learning algorithm model can accurately predict the molecular composition data of unknown crude oil samples.

[0022] The molecular composition analysis model outputs data of the crude oil sample through the machine learning algorithm model, and compares it with the actual molecular composition data of the crude oil sample to verify the prediction accuracy of the machine learning algorithm model. If the error of the predicted molecular composition data is greater than a preset threshold, the parameters of the model are adjusted until the error is less than or equal to the preset threshold, thus obtaining the molecular composition analysis model.

[0023] In one embodiment, the machine learning algorithm model includes one or more of the following models: support vector machine, neural network, and random forest algorithm.

[0024] In one embodiment, the above-mentioned oil refining simulation system further includes: a molecular-level reaction rule database;

[0025] The molecular reaction rule database is used to provide information on various reaction rules and corresponding kinetic parameters for the molecular composition analysis model and the secondary processing device model.

[0026] In one embodiment, the molecular reaction rule database is generated in advance by the following method:

[0027] The existing chemical reaction rules data in the refining and chemical industry are classified according to the type of raw material molecules;

[0028] For each type of raw material molecule, the corresponding reaction rules are used to obtain the kinetic parameters of the reaction rules, including activation energy and reaction rate constant.

[0029] The types, reaction rules, and kinetic parameters of each raw material molecule are stored to generate the molecular reaction rule database.

[0030] Secondly, embodiments of the present invention provide a method for simulating molecular-level reactive mass flow in a refining process using the aforementioned refining simulation system, the method comprising the following steps:

[0031] Receive physical property data and chemical analysis characteristic data of crude oil samples input by users, use a trained molecular composition analysis model to predict and output the molecular composition and content of the crude oil samples;

[0032] Based on the predicted molecular composition and content of the crude oil sample, the reaction process of crude oil in the secondary processing unit is simulated to predict the molecular composition and physical properties of the reaction products.

[0033] Based on the target physical property parameters of the blended product, as well as the predicted molecular composition and physical property parameters of the reaction products, the optimal blending ratio to meet the target physical property parameters of the blended product is calculated, and the blending operation is simulated using the optimal blending ratio.

[0034] Thirdly, embodiments of the present invention provide a computing device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for simulating molecular-level reactive mass flow in an oil refining process as described above.

[0035] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for simulating molecular-level reactive mass flow in an oil refining process as described above.

[0036] Fifthly, an embodiment of the present invention provides a computer program product, the computer program product comprising a computer program, which, when executed by a processor, implements the method for simulating the molecular-level reactive mass flow of the oil refining process as described above.

[0037] The beneficial effects of the above-described technical solutions provided in the embodiments of the present invention include at least the following:

[0038] The oil refining simulation system provided in this invention includes three models: a molecular composition analysis model, a secondary processing unit model, and a blending optimization model. It simulates the molecular composition of crude oil samples at the molecular level. Based on the crude oil molecular composition, the secondary processing unit model simulates the reaction process, predicting the molecular composition and content of reaction products under specific processing conditions. The blending optimization model calculates the optimal blending ratio that meets the target product's physical property parameters based on the molecular composition and content of the reaction products output by the secondary processing unit model and simulates the blending process. This oil refining simulation system provides a complete model of the oil refining process at the molecular level, enabling accurate simulation at this stage. This improves the adaptability and efficiency of the oil refining process and allows users to intuitively understand molecular-level changes during refining, providing a powerful analytical tool for optimizing refining processes.

[0039] 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.

[0040] 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

[0041] 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:

[0042] Figure 1 , 2 This is a block diagram of the oil refining simulation system in an embodiment of the present invention;

[0043] Figure 3 This is a flowchart of a method for simulating molecular-level reactive mass flow in an oil refining process, as described in an embodiment of the present invention. Detailed Implementation

[0044] 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.

[0045] To address the problems in the prior art, this invention provides a refining simulation system that can accurately simulate the refining process at the molecular level, thereby improving the adaptability and efficiency of the refining process.

[0046] In practice, this oil refining simulation system can be implemented using a software system. The components and functions of the oil refining system are described in detail below.

[0047] An oil refining simulation system provided in this embodiment of the invention, with reference to Figure 1 As shown, it includes: a molecular composition analytical model, a secondary processing device model, and a harmonic optimization model; wherein:

[0048] The molecular composition analysis model is used to receive the physical property data and chemical analysis characteristic data of the input crude oil sample, and use the trained machine learning algorithm model to predict the molecular composition and content of the crude oil sample and output it to the secondary processing device model.

[0049] The secondary processing unit model is used to simulate the reaction process of crude oil in the secondary processing unit based on the molecular composition and content of the crude oil sample output by the molecular composition analysis model, predict the molecular composition and physical property parameters of the reaction products, and output them to the harmonization optimization model.

[0050] The blending optimization model is used to calculate and iteratively adjust the blending ratio of the blended product based on the molecular composition and physical property parameters output by the secondary processing device model and the preset target physical property parameters of the blended product, until it meets the preset target physical property parameters, obtain the optimal value of the blending ratio, and use the optimal value of the blending ratio to simulate the blending operation.

[0051] The oil refining simulation system provided in this invention includes three models: a molecular composition analysis model, a secondary processing unit model, and a blending optimization model. It simulates the molecular composition of crude oil samples at the molecular level. Based on the crude oil molecular composition, the secondary processing unit model simulates the reaction process, predicting the molecular composition and content of reaction products under specific processing conditions. The blending optimization model calculates the optimal blending ratio that meets the target product's physical property parameters based on the molecular composition and content of the reaction products output by the secondary processing unit model and simulates the blending process. This oil refining simulation system provides a complete model of the oil refining process at the molecular level, enabling accurate simulation at this level. It improves the adaptability and efficiency of the oil refining process and allows users to intuitively understand molecular-level changes during refining, providing a powerful analytical tool for optimizing refining processes.

[0052] In one embodiment, the above-mentioned molecular composition analysis model can be pre-constructed and trained in the following manner:

[0053] 1.1 Determine the molecular structure and content of each component in the crude oil sample to obtain data on the molecular composition of the crude oil sample;

[0054] 1.2. Characteristic data of the molecular composition of crude oil samples were obtained through chemical analysis;

[0055] These characteristic data may include, for example, the peak area, retention time, peak height, and peak shape parameters of the compounds corresponding to each component;

[0056] 1.3 The data on the molecular composition of the crude oil sample, the feature data corresponding to the molecular composition, and the physical property data are used to form a dataset for model training;

[0057] 1.4 Using the dataset, train the pre-architected machine learning algorithm model. During the training process, adjust the parameters of the machine learning algorithm model so that it can accurately predict the molecular composition data of unknown crude oil samples.

[0058] 1.5. The molecular composition analysis model outputs the molecular composition data of the known crude oil sample through the machine learning algorithm model, and compares it with the molecular composition data of the known crude oil sample to verify the prediction accuracy of the output of the machine learning algorithm model. If the error of the predicted molecular composition data is greater than a preset threshold, the parameters of the model are adjusted until the error is less than or equal to the preset threshold, and the molecular composition analysis model is obtained.

[0059] Optionally, the determination of the structure and content of the crude oil sample in steps 1.1-1.2 above, as well as the chemical analysis characteristic data obtained during the chemical analysis process, can be obtained, for example, through gas chromatography-mass spectrometry (GC-MS). GC-MS is an instrumental analytical method that combines gas chromatography (GC) and mass spectrometry (MS). It can be used to identify and determine the molecular composition and content of each component in complex mixtures.

[0060] During GC-MS analysis, chemical analysis characteristic data can be obtained simultaneously, such as peak area, retention time, peak height, and peak shape parameters of the corresponding compounds for each component.

[0061] After obtaining GC-MS data for chemical analysis using GC-MS technology, filtering algorithms (such as Savitzky-Golay filtering) can be applied to reduce random noise in the data, improve the signal-to-noise ratio, and make subsequent peak identification more accurate.

[0062] The software then automatically identifies the peaks in the chromatogram. This process may involve setting parameters such as peak height and peak width to distinguish between true peaks and noise.

[0063] After peak identification, the identified chromatographic peaks are integrated to calculate their peak area, peak height, and relevant parameters defining the peak shape. The peak area integration results provide the relative content information of each component for the blending optimization model, used to calculate the optimal blending ratio. The retention times obtained from the experiment are recorded.

[0064] In this embodiment of the invention, the definitions of the peak area, retention time, peak height, and peak shape parameters are explained as follows:

[0065] Peak area refers to the area of ​​each peak calculated using an integral method, representing the relative content of the compound.

[0066] Retention time refers to the retention time of each peak, which is the residence time of a compound in the chromatographic column. Specifically, it refers to the time it takes for a compound in the sample to be detected by the mass spectrometer after passing through the chromatographic column, starting from the injection port of the gas chromatograph. The unit is usually minutes.

[0067] Peak height refers to the height value of each extracted peak, reflecting the response intensity of the compound.

[0068] Peak shape parameters include: symmetry parameters used to calculate peak symmetry (such as symmetry factor), peak width (including half-width at half-maximum), etc.

[0069] The determination of the molecular structure and content of crude oil by GC-MS technology can be carried out by referring to the operation methods in the prior art, and will not be repeated in the embodiments of this invention.

[0070] In step 1.3 above, the training set can consist of the following three parts:

[0071] ① Input features: The standardized numerical vector of crude oil physical property data (such as density, viscosity, boiling point);

[0072] ② Correlation features: chromatographic feature vectors such as peak area, retention time, peak height, and peak shape parameters analyzed by GC-MS;

[0073] ③ Output labels: A vector of molecular composition proportions verified by mass spectrometry.

[0074] In practical implementation, when using a trained molecular composition analysis model for prediction, for input data containing only physical property parameters, it is necessary to first establish a correlation mapping matrix between physical property parameters and chromatographic features, and then input the completed features into the machine learning algorithm model for molecular composition prediction.

[0075] The selected machine learning algorithm model is trained using the integrated dataset described above, and the model parameters are adjusted to obtain the best predictive performance.

[0076] The selected machine learning algorithm model is trained using an integrated dataset, and the model parameters are tuned to obtain the best predictive performance.

[0077] The model's predictive accuracy was validated by comparison with an independent test dataset containing crude oil samples with known molecular compositions.

[0078] The trained model is applied to the analysis of actual crude oil samples to predict their molecular composition, providing input data for subsequent refining processes.

[0079] Optionally, the aforementioned machine learning algorithm model may include one or more of the following models: support vector machine, neural network, and random forest algorithm, etc.

[0080] Existing GC-MS analysis technology uses data such as peak area, peak height, retention time, and peak shape parameters to identify the chemical structure of each component. Because these chemical analysis feature data obtained through chemical analysis have a close intrinsic relationship (mapping relationship) with the molecular composition of crude oil, the molecular composition and content of crude oil can be well identified. Therefore, the combination of physical property data and chemical analysis feature data can provide good data support for training a molecular composition analysis model that can accurately predict the molecular composition information of crude oil, thus improving the robustness of the model. For the molecular composition analysis model provided in the embodiments of this invention, it is precisely by using GC-MS data to establish the mapping relationship between crude oil physical property data, chemical analysis feature data, and molecular composition and content that accurate raw material molecular composition information is provided for the refining process.

[0081] In one embodiment, the above-mentioned secondary processing device model is pre-built and trained in the following manner:

[0082] 2.1 Constructing the molecular composition matrix of raw materials;

[0083] The molecular composition matrix is ​​a matrix that describes the composition of each molecule in the raw material. Each row represents a type of molecule, and each column represents the number of atoms of different elements in the molecule.

[0084] 2.2 Based on the molecular composition matrix of the raw materials, and combined with the reaction rules and kinetic equations of catalytic cracking, a mathematical model is constructed to describe the relationship between the reaction rate and the concentrations of reactants and products under different reaction rules in the secondary processing unit.

[0085] 2.3 By comparing the actual products corresponding to the raw materials with the raw material molecular composition matrix, the kinetic parameters in the mathematical model are continuously adjusted to minimize the difference between the predicted products and the actual products until the preset accuracy requirements are met.

[0086] Specifically, the above-mentioned harmonic optimization model can be pre-built and optimized, for example, in the following ways:

[0087] Based on the molecular composition and physical property parameters output by the secondary processing device model, the initial value of the blending ratio is calculated using an optimized mathematical algorithm.

[0088] Based on the initial blending ratio, the physical properties of the blended product are predicted. The blending result is evaluated by comparing the predicted physical properties of the blended product with the preset target physical properties. If the difference is greater than the preset difference threshold, the blending ratio is adjusted, and the physical properties of the blended product are predicted again based on the adjusted blending ratio. The blending result is evaluated again based on the difference between the predicted physical properties of the blended product and the target physical properties, until the difference is lower than the preset difference threshold, thus obtaining the optimal blending ratio.

[0089] Based on the target physical property parameters, the harmonic ratio is continuously calculated by optimizing the mathematical algorithm. This ensures that the optimal harmonic ratio obtained after optimization is as close as possible to the target physical property parameters, guaranteeing that the product obtained by the harmonic optimization model through simulation meets the aforementioned target physical property parameters.

[0090] In one embodiment, the initial value of the harmonic optimization model described above can be calculated specifically by the following method:

[0091] Using a mathematical optimization algorithm, based on the molecular composition and physical property parameters of the reaction products output by the secondary processing device model, a system is constructed with the objective function of minimizing the weighted square difference between the target physical property parameters (e.g., octane number, density, sulfur content, etc.) and the physical property parameters output by the secondary processing device model. Combining the physical property parameters and content constraints of each component, the optimal solution is obtained through numerical solution to obtain the initial value of the blending ratio.

[0092] The adjustment of the blending ratio can be mainly followed according to the following principles:

[0093] (1) Principle of matching physical properties: Prioritize meeting the requirements of key physical property indicators such as octane number and density.

[0094] (2) Economic principle: Under the premise of meeting quality standards, priority should be given to the high proportion of low-cost components.

[0095] The aforementioned mathematical optimization algorithm can be implemented using various constrained optimization algorithms, such as linear programming, sequential quadratic programming, Lagrange multiplier method, etc. The specific mathematical optimization algorithm used in the embodiments of this invention is not limited.

[0096] In one embodiment, the above-mentioned oil refining simulation system refers to Figure 2 As shown, it may also include: a molecular-level reaction rule database; this molecular reaction rule database is used to provide information on various reaction rules and corresponding kinetic parameters for the molecular composition analytical model and the secondary processing device model.

[0097] In one embodiment, the molecular reaction rule database can be generated in advance, for example, in the following manner:

[0098] The existing chemical reaction rules data in the refining and chemical industry are classified according to the type of raw material molecules;

[0099] For each type of raw material molecule, the corresponding reaction rules are used to obtain the kinetic parameters of the reaction rules, including activation energy and reaction rate constant.

[0100] The types, reaction rules, and kinetic parameters of each raw material molecule are stored to generate the molecular reaction rule database.

[0101] For example, existing chemical reaction rules and data in the refining and chemical industry can be obtained by collecting and organizing existing chemical reaction data, including literature, experimental reports, and industrial data.

[0102] The above-mentioned reaction rules can be classified according to the type of raw material molecules, such as aromatic hydrocarbons, cycloalkanes, alkenes, alkanes, and molecules containing heteroatoms.

[0103] For each type of raw material molecule, the database contains reaction rules that may occur under different reaction conditions, as well as corresponding reaction kinetic parameters, such as activation energy and reaction rate constant.

[0104] With the accumulation of new chemical reaction data and a deeper understanding of reaction mechanisms, the molecular-level reaction rule database needs to be continuously updated and improved to adapt to the development of new raw material characteristics and processing technologies, and to ensure the accuracy of the output results of the above-mentioned molecular composition analysis model and secondary processing device model.

[0105] The molecular reaction rule database in this embodiment of the invention contains reaction rules for various raw material molecules, covering a wide range of reaction types and conditions, providing comprehensive support for the model. By collecting and organizing reliable chemical reaction data, the information in the database has high accuracy, which helps to improve the predictive performance of the model. The database can provide customized reaction rules and kinetic parameters for the model according to different raw material characteristics and reaction conditions, thereby improving the adaptability of the model.

[0106] In one embodiment, the aforementioned oil refining simulation system can be implemented using molecular oil refining simulation software (hereinafter referred to as the software). This software tracks the transformation paths of raw material molecules during the oil refining process through simulation calculations, predicting the molecular composition and physical properties of the products. This helps optimize the oil refining process and improve product quality and production efficiency.

[0107] The working process of molecular refining simulation software can be briefly described as follows: The software first receives crude oil physical property data and molecular composition information input by the user. This data can come from laboratory tests, field measurements, or historical data.

[0108] Based on the crude oil physical properties and molecular composition input by the user, the software calls upon a molecular composition analysis model for analysis. This model uses machine learning algorithms to predict the molecular structure and content of each component in the crude oil.

[0109] Using the analysis results from the molecular composition analytical model, the software further calls upon a secondary processing unit model to simulate the reaction process of crude oil in specific processing units (such as catalytic cracking and hydrocracking). This model predicts the molecular composition of the reaction products by adjusting kinetic parameters.

[0110] Next, we will further use simulation calculations to trace the transformation pathways of feedstock molecules during the refining process. This includes the consumption, generation, and transformation of feedstock molecules at different reaction stages.

[0111] The software summarizes the content of raw material and product molecules for all effective reaction pathways, generating a product molecule composition matrix for secondary processing reactions. This matrix provides the foundational data for subsequent harmonic optimization models.

[0112] Based on the product molecular composition matrix, the software predicts the physical properties of the blended product, such as octane number, boiling point, and density. These parameters are crucial for assessing whether the blended product meets specific quality standards.

[0113] Based on the predicted physical property parameters, the software invokes a blending optimization model to calculate the optimal blending ratio. Through iterative adjustments, it finds the optimal blending ratio that ensures the physical property parameters of the blended product meet the preset target.

[0114] The aforementioned molecular refining simulation software can simulate the molecular-level reaction material flow in the refining process, providing comprehensive analysis results. Through predictions using molecular composition analysis models and secondary processing unit models, the software can accurately track the transformation path of raw material molecules and predict the molecular composition of products. It can flexibly adjust the simulation calculation process based on different crude oil physical property data and molecular composition input by the user, adapting to the changing refining process requirements. Through the blending optimization model, it provides refining enterprises with an effective tool for optimizing blended product physical property parameters, which helps improve product quality and production efficiency.

[0115] Based on the same inventive concept, this embodiment of the invention also provides a method for simulating molecular-level reaction material flow in the refining process using the aforementioned refining simulation system. Since the principle of the problem solved by this method is similar to that of the aforementioned refining simulation system, the implementation of each step of this method can refer to the implementation of the aforementioned refining simulation system, and the repeated parts will not be described again.

[0116] The method for simulating molecular-level reactive mass flow in the refining process using the aforementioned refining simulation system, as provided in this embodiment of the invention, refers to... Figure 3 As shown, the method includes the following steps:

[0117] S31. Receive the physical property data and chemical analysis characteristic data of the crude oil sample input by the user, and use the trained molecular composition analysis model to predict and output the molecular composition and content of the crude oil sample.

[0118] The chemical analysis characteristic data mentioned above can also be obtained by GC-MS analysis, similar to the aforementioned embodiments.

[0119] S32. Based on the predicted molecular composition and content of the crude oil sample, simulate the reaction process of crude oil in the secondary processing unit to predict the molecular composition and physical property parameters of the reaction products.

[0120] S33. Based on the target physical property parameters of the blended product and the predicted molecular composition and physical property parameters of the reaction products, calculate the optimal blending ratio to meet the target physical property parameters of the blended product, and use the optimal blending ratio to simulate the blending operation.

[0121] The specific implementation methods of steps S31-S33 above can be referred to the specific implementation methods of the above embodiments, and the repeated parts will not be described again.

[0122] This invention also provides a computing 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 method for simulating molecular-level reactive mass flow in an oil refining process as provided in the foregoing embodiments.

[0123] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for simulating molecular-level reactive mass flow in an oil refining process as described above.

[0124] The computer program product provided in this embodiment of the invention includes a computer program, which, when executed by a processor, is a method for simulating molecular-level reactive mass flow in an oil refining process as described above.

[0125] 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.

[0126] 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.

[0127] 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.

[0128] 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.

[0129] 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 refining simulation system, characterized in that, include: Molecular composition analytical model, secondary processing device model, and harmonic optimization model; among which: The molecular composition analysis model is used to receive the physical property data and chemical analysis characteristic data of the input crude oil sample, and use the trained machine learning algorithm model to predict the molecular composition and content of the crude oil sample and output it to the secondary processing device model. The secondary processing device model is used to simulate the reaction process of the crude oil sample in the secondary processing device based on the molecular composition and content of the crude oil sample output by the molecular composition analysis model, predict the molecular composition and physical property parameters of the reaction products, and output them to the harmonization optimization model. The blending optimization model is used to calculate and iteratively adjust the blending ratio of the blended product based on the molecular composition and physical property parameters output by the secondary processing device model and the preset target physical property parameters of the blended product, until it meets the preset target physical property parameters, obtain the optimal value of the blending ratio, and use the optimal value of the blending ratio to simulate the blending operation.

2. The oil refining simulation system as described in claim 1, characterized in that, The harmonic optimization model is pre-built and optimized in the following manner: Based on the molecular composition and physical property parameters output by the secondary processing device model, the initial value of the blending ratio is calculated using an optimized mathematical algorithm. Based on the initial blending ratio, the physical properties of the blended product are predicted. The blending result is evaluated by comparing the predicted physical properties of the blended product with the preset target physical properties. If the difference is greater than the preset difference threshold, the blending ratio is adjusted, and the physical properties of the blended product are predicted again based on the adjusted blending ratio. The blending result is evaluated again based on the difference between the predicted physical properties of the blended product and the target physical properties, until the difference is lower than the preset difference threshold, thus obtaining the optimal blending ratio.

3. The oil refining simulation system as described in claim 2, characterized in that, The aforementioned harmonization optimization model specifically employs a mathematical optimization algorithm. Based on the molecular composition and physical property parameters of the reaction products output by the secondary processing device model, it constructs a model with the objective function of minimizing the weighted square difference between the target physical property parameters and the physical property parameters output by the secondary processing device model. Combining the physical property parameters and content constraints of each component, the model numerically solves for the optimal solution to obtain the initial value of the harmonization ratio.

4. The oil refining simulation system as described in claim 1, characterized in that, The secondary processing device model was pre-built and trained in the following manner: Construct a molecular composition matrix for the raw materials; the molecular composition matrix is ​​a matrix describing the composition of each molecule in the raw materials, with each row representing a type of molecule and each column representing the number of atoms of different elements in the molecule; Based on the molecular composition matrix of the raw materials, and combined with the reaction rules and kinetic equations of catalytic cracking, a mathematical model is constructed to describe the relationship between the reaction rate and the concentrations of reactants and products under different reaction rules in the secondary processing unit. By comparing the actual products corresponding to the raw materials with the raw material molecular composition matrix, the kinetic parameters in the mathematical model are continuously adjusted to minimize the difference between the predicted products and the actual products until the preset accuracy requirements are met.

5. The oil refining simulation system as described in claim 1, characterized in that, The molecular composition analytical model was pre-constructed and trained in the following manner: The molecular structure and content of each component in the crude oil sample were determined to obtain data on the molecular composition of the crude oil sample. Characteristic data of the molecular composition of crude oil samples were obtained through chemical analysis; The data on the molecular composition, molecular composition features, and physical properties of the crude oil sample are used to form a dataset for model training. Using the dataset, a pre-architected machine learning algorithm model is trained. During the training process, the parameters of the machine learning algorithm model are adjusted so that the machine learning algorithm model can accurately predict the molecular composition data of unknown crude oil samples. The molecular composition analysis model outputs data of the crude oil sample through the machine learning algorithm model, and compares it with the actual molecular composition data of the crude oil sample to verify the prediction accuracy of the machine learning algorithm model. If the error of the predicted molecular composition data is greater than a preset threshold, the parameters of the model are adjusted until the error is less than or equal to the preset threshold, thus obtaining the molecular composition analysis model.

6. The oil refining simulation system as described in claim 5, characterized in that, The machine learning algorithm model includes one or more of the following models: support vector machine, neural network, and random forest algorithm.

7. The oil refining simulation system according to any one of claims 1-6, characterized in that, Also includes: Molecular-level reaction rule database; The molecular reaction rule database is used to provide information on various reaction rules and corresponding kinetic parameters for the molecular composition analysis model and the secondary processing device model.

8. The oil refining simulation system as described in claim 7, characterized in that, The molecular reaction rule database is generated in advance using the following method: The existing chemical reaction rules data in the refining and chemical industry are classified according to the type of raw material molecules; For each type of raw material molecule, the corresponding reaction rules are used to obtain the kinetic parameters of the reaction rules, including activation energy and reaction rate constant. The types, reaction rules, and kinetic parameters of each raw material molecule are stored to generate the molecular reaction rule database.

9. A method for simulating molecular-level reactive mass flow in a refining process using a refining simulation system as described in any one of claims 1-8, characterized in that, The method includes the following steps: Receive physical property data and chemical analysis characteristic data of crude oil samples input by users, use a trained molecular composition analysis model to predict and output the molecular composition and content of the crude oil samples; Based on the predicted molecular composition and content of the crude oil sample, the reaction process of crude oil in the secondary processing unit is simulated to predict the molecular composition and physical properties of the reaction products. Based on the target physical property parameters of the blended product, as well as the predicted molecular composition and physical property parameters of the reaction products, the optimal blending ratio to meet the target physical property parameters of the blended product is calculated, and the blending operation is simulated using the optimal blending ratio.

10. A computing 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 simulating molecular-level reactive mass flow in a refining process as described in claim 9.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method for simulating molecular-level reactive mass flow in a refining process as described in claim 9.

12. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method for simulating molecular-level reactive mass flow in a refining process as described in claim 9.

Images