Method and system for automatically generating reaction rule of molecular level model

Abstract

The invention provides a reaction rule automatic generation method and system of a molecular level model, and relates to the technical field of technological process simulation. An initial rule is constructed based on a basic reaction type in the oil refining field to generate an initial reaction rule set; determining screened reactants from the actual reactant set, and determining generated products from the actual product set; determining a first reactant set according to the screened reactants, the actual reactant set and the actual non-reaction reactant set; adding a reaction rule matched with each first reactant into the initialized reaction rule set according to the structural characteristics of the groups; determining a second reactant set according to the generated product, the actual product set and the actual reactant set; and obtaining a target rule set according to the reaction rule matched with each second reactant according to the group structure characteristics. According to the reaction rule generation method, omission reaction rules can be automatically identified, professional intervention is not needed, the modeling threshold is lowered, and the modeling efficiency and precision are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of process simulation, in particular to a reaction rule automatic generation method and system of a molecular level model. BACKGROUND

[0002] Process simulation is a mathematical expression of the chemical thermodynamics and kinetics of the actual refinery conversion process using computer technology, aiming to reproduce the component movement and conversion law through the model, and then diagnose, predict and guide the actual operation of the device. With the increasing demand for fine production, traditional virtual component or lumped component modeling has been difficult to meet the accuracy requirements, and molecular level modeling has become an inevitable trend.

[0003] CN111899812A develops a product simulation method for a petroleum processing device, which models the process to the molecular level. The model is constructed in four steps: feedstock input, reaction rule setting, reactor configuration, and kinetic parameter parameterization. The setting of the reaction rule needs to reflect the actual chemical reaction process, and its setting directly affects the model output effect, which is a key step in molecular level modeling. CN115841852B provides a method and device for determining the reaction rule of catalytic reforming, which describes a method for determining the reaction rule of a catalytic reforming device. The reactants are screened and determined, but the reactions that actually occur but may be missed by the rule set are not described, and the product generation rule is not described in detail. The requirement for the modeling personnel's background in oil refining and chemical industry is high. CN115831248A also provides a method and device for determining a reaction rule, an electronic device and a storage medium, and CN115841851A provides a method and device for constructing a molecular level reaction rule for hydrocracking. However, in the prior art, the reaction rule is supplemented by professional personnel manually. The missing reactions (such as unreacted reactants and ungenerated products) in the model need to be identified based on experience, and the rules are added manually, which is not only inefficient, but also requires a high level of background in oil refining and chemical industry for modeling personnel, making it difficult for non-professionals to operate and limiting the popularization and application of molecular level models. In addition, manual rule supplementation is prone to omission or misjudgment, affecting the stability of model accuracy.

[0004] Therefore, there is an urgent need for a reaction rule generation method that can automatically identify missing reaction rules without the intervention of professional personnel, in order to reduce the modeling threshold and improve modeling efficiency and accuracy. SUMMARY

[0005] The embodiments of the present application provide a reaction rule automatic generation method and system of a molecular level model, which can effectively reduce the operation threshold and improve the efficiency compared to the manual supplementation of reaction rules in the prior art molecular level model.

[0006] To achieve the above purpose, the embodiments of the present application adopt the following technical solutions: In a first aspect, a method for automatically generating reaction rules of a molecular level model is provided. The method comprises: obtaining actual feedstock data and actual product data of a refinery to be modeled, performing data set classification processing to obtain actual reactant sets, actual product sets and actual non-reactant sets at a molecular level; wherein the actual reactant sets include molecules that can participate in chemical reactions in the refinery, the actual product sets include molecules generated in the actual production process, and the actual non-reactant sets include molecules that cannot participate in chemical reactions in the refinery; constructing initial rules based on basic reaction types in the field of the refinery, combining the actual reactant sets and the actual product sets to generate an initial reaction rule set; the initial reaction rule set includes a plurality of reaction rules, and each reaction rule includes reactant screening conditions and product generation conditions; determining screened reactants from the actual reactant sets and determining generated products from the actual product sets, wherein the screened reactants satisfy the reactant screening conditions included in the initial reaction rule set, and the generated products satisfy the product generation conditions included in the initial reaction rule set; determining a first reactant set according to the screened reactants, the actual reactant sets and the actual non-reactant sets; wherein the first reactant set includes a plurality of first reactants and group structure characteristics of each first reactant, and the first reactants are molecules that belong to the actual reactant sets, do not belong to the screened reactants, and do not belong to the actual non-reactant sets; matching each first reactant with a plurality of reaction rules included in a preset reaction rule set according to the group structure characteristics of each first reactant, adding the matched reaction rules of each first reactant to the initial reaction rule set to obtain an updated rule set, and the preset reaction rule set is a standardized rule set that is pre-constructed and verified by actual reactions in the field of the refinery; determining a second reactant set according to the generated products, the actual product sets and the actual reactant sets; wherein the second reactant set includes a plurality of second reactants and group structure characteristics of each second reactant, and the second reactants are molecules that belong to the actual reactant sets and generate products that belong to the actual product sets but do not belong to the generated products; matching each second reactant with a plurality of reaction rules included in the preset reaction rule set according to the group structure characteristics of the second reactants, and adding the matched reaction rules of each second reactant to the updated rule set until the first reactant set and the second reactant set are empty to obtain a target rule set.

[0007] The method provided by the present application solves the problem of manual construction of reaction rules in the traditional molecular level model by the whole process design of data acquisition and classification, initial rule set construction, screening of reaction products, first / second reactant set supplement, rule iteration and update until the reactant set is empty. In the traditional method, manual rule construction is not only inefficient, but also prone to incomplete rule coverage due to the limitations of artificial experience, such as missing some reaction molecules or corresponding rules for actual product generation. The method provided by the present application can first obtain accurate reaction product sets from actual data through an automatic process, and then identify reaction molecules that are not covered by the initial rules through the first reaction set, and trace back the reaction molecules that do not generate products through the second reaction set. Finally, the rules are iteratively completed until there is no omission, and the automatic and comprehensive generation of reaction rules is finally realized. The method avoids the subjectivity and limitations of manual construction, significantly improves the efficiency and completeness of rule generation, and meets the core needs of the molecular level model for accurate and comprehensive reaction rules.

[0008] In a possible implementation manner of the first aspect, the actual feedstock data at least includes structure-oriented lumped characteristics and mass fractions of each molecule in the feedstock of the to-be-modeled oil refining device, and the actual product data at least includes structure-oriented lumped characteristics and mass fractions of each molecule in the actual production product of the to-be-modeled oil refining device; the data set classification processing adopts a structure-oriented lumped classification algorithm, and by identifying whether the structure-oriented lumped characteristics of the molecule contain active groups that can participate in oil refining reactions, the molecules containing active groups in the actual feedstock data are divided into an actual reactant set, and the inert molecules not containing active groups are divided into an actual non-reactant set; at the same time, by comparing the structure-oriented lumped characteristics of the molecule with the actual production record, the molecules generated in the actual production process are screened into an actual product set.

[0009] The method provided by the present application determines that the actual feedstock and product data need to include structure-oriented lumped (SOL) characteristics and mass fractions, and limits the use of a structure-oriented lumped classification algorithm for data processing, and classifies the reaction products by identifying active groups and screening the products by comparing production records. In the method provided by the present application, the structure-oriented lumped characteristics can accurately reflect the core reaction structure of the molecule, the mass fraction provides a basis for data quantization, and the structure-oriented lumped classification algorithm ensures that only molecules containing active groups are classified as reaction products, and only molecules in the actual production record are classified as products, thereby significantly improving the accuracy of data processing and ensuring the authenticity and effectiveness of the actual reaction product set and the product set, providing a reliable data basis for subsequent initialization rule construction, screening matching and other steps, and reducing the rule deviation caused by inaccurate data.

[0010] In a possible implementation manner of the first aspect, the basic reaction types in the oil refining field include a six-membered ring dehydroaromatization reaction and a five-membered ring isomerization reaction, the reactant screening condition of the six-membered ring dehydroaromatization reaction is that a molecule contains a six-membered ring structure feature, the reactant screening condition of the five-membered ring isomerization reaction is that a molecule contains a five-membered ring structure feature, the product generation condition of the six-membered ring dehydroaromatization reaction is to generate a molecule containing a monocyclic aromatic hydrocarbon structure feature, and the product generation condition of the five-membered ring isomerization reaction is to generate a molecule containing a five-membered ring derivative structure feature.

[0011] The method provided by the present application clearly defines the basic reaction types in the oil refining field as the six-membered ring dehydroaromatization reaction and the five-membered ring isomerization reaction, and directly and correspondingly defines the reactant screening conditions and the product generation conditions of the two types of reactions. The method provided by the present application focuses on the core ring reaction in the oil refining field, and directly binds the reaction types and the molecule SOL features, so that the screening conditions and the generation conditions of the initial rules are visualized and can be implemented, the ambiguity of the initial rules is avoided, clear and executable bases are provided for the determination of the screened reactants or the generated products, and the initial rules can accurately cover the key basic reactions in the oil refining field, thereby laying a clear starting point for rule iteration.

[0012] In a possible implementation manner of the first aspect, the determining of the screened reactants from the actual reactant set comprises: calling a preset reactant screening program, the reactant screening program loads the reactant screening condition corresponding to each reaction rule in the initialized reaction rule set, and performs structure-oriented collective feature matching on each molecule in the actual reactant set, and molecules that pass the matching are determined as the screened reactants; and the determining of the generated products from the actual product set comprises: calling a preset product generation program, the product generation program simulates the molecular conversion process of the screened reactants based on the product generation condition corresponding to each reaction rule in the initialized reaction rule set, generates simulation products, and then performs structure-oriented collective feature comparison on the simulation products and the actual product set, and the simulation products that exist in the actual product set are determined as the generated products.

[0013] In the method provided by the present application, the preset program automatically matches the SOL features, can quickly and accurately screen the reactants that meet the rules, the product generation program verifies the simulation conversion and the actual comparison, ensures that the generated products are molecules that exist in actual production, reduces the errors caused by manual intervention, greatly improves the efficiency of the screening and generation processes, and provides reliable intermediate results for the identification of the first / second reactant set, thereby avoiding the deviation of the entire rule generation process caused by the errors of the intermediate results.

[0014] In a possible implementation manner of the first aspect, the determining the first reactant set according to the screened reactants, the actual reactant set and the actual non-reactant set comprises: comparing the screened reactants with the actual reactant set, extracting molecules in the actual reactant set that are not included in the screened reactants to obtain a non-screened molecule list; comparing the non-screened molecule list with the actual non-reactant set, eliminating inert molecules in the non-screened molecule list that belong to the actual non-reactant set, and the remaining molecules constitute the first reactants; performing structure-oriented collective feature extraction on each first reactant to obtain group structure features of each first reactant, and determining the first reactant set according to the group structure features of each first reactant.

[0015] In the method provided by the application, the first step accurately locates molecules that are not covered by the initial rules by comparing the screened reactants with the actual reactant set; the second step eliminates non-reactive molecules to ensure that the remaining molecules are all effective molecules that can react but are not covered; and the third step extracts group structure features to provide clear targeting information for subsequent rule matching and rule completion. This process can accurately identify the omissions of the initial rules, avoid the lack of rules caused by the non-coverage of the reactive molecules, and provide clear basis for rule completion, thereby ensuring the pertinence of subsequent rule iteration.

[0016] In a possible implementation manner of the first aspect, the preset reaction rule set comprises alkyl isomerization reaction rules, alkyl dehydrogenation cyclization reaction rules, alkyl hydrogenation cracking reaction rules, five-membered ring isomerization reaction rules, six-membered ring dehydrogenation aromatization reaction rules, hydrogenation desulfurization reaction rules, hydrogenation denitrification reaction rules, and hydrogenation deoxygenation reaction rules; each type of reaction rule includes group matching conditions and a molecular conversion path, wherein the group matching conditions are constraints corresponding to the group structure features of the first reactants, and the molecular conversion path is a standard oil refining chemical reaction path corresponding to the group.

[0017] The eight types of rules covered by the method provided by the application cover main reaction types in the oil refining field, from alkyl reactions, ring reactions to heteroatom removal, ensuring comprehensive rule resources; the group matching conditions provide clear standards for matching the first reactants with the rules, and the molecular conversion path ensures that the rules conform to the oil refining reaction rules, thereby avoiding randomness in rule matching and improving the effectiveness of rule completion, and providing reliable and standardized resource support for rule completion of the first reactants.

[0018] In a possible implementation manner of the first aspect, the matching of the group structure feature of each first reactant with the plurality of reaction rules contained in the preset reaction rule set, the adding of the matched reaction rule of each first reactant into the initial reaction rule set to obtain the updated rule set, comprises: analyzing the group structure feature of each first reactant to determine a core reaction group of each first reactant; searching for a reaction rule in the preset reaction rule set, which has a group matching condition completely corresponding to the core reaction group; if the corresponding reaction rule is found and does not exist in the initial reaction rule set, the corresponding reaction rule is added into the initial reaction rule set to complete rule supplementing, and the updated rule set is formed.

[0019] In the method provided by the application, the analysis of the core reaction group ensures that the matched rule is for the key reaction site of the first reactant, and avoids the rule from being irrelevant to the reactant; the searching for the matched rule ensures that the rule meets the preset standard; and the adding of only the rule that does not exist avoids repeated adding and keeps the rule set simple. The flow realizes accurate and non-redundant supplementing of the rule, ensures that the supplemented rule can accurately cover the reaction corresponding to the first reactant, avoids the subsequent execution efficiency from being reduced due to the bloated rule set, and improves the pertinence and practicality of the updated rule set.

[0020] In a possible implementation manner of the first aspect, the determining of the second reactant set according to the generated product, the actual product set and the actual reactant set comprises: comparing the generated product with the actual product set to extract molecules in the actual product set that are not included in the generated product, to obtain an ungenerated product list; based on the molecular conversion rule in the refining field, tracing precursor molecules in the actual reactant set that can generate each molecule in the ungenerated product list; performing structure-oriented collective feature extraction on each precursor molecule to obtain a group structure feature of each precursor molecule, and determining the second reactant set according to the group structure feature of each precursor molecule.

[0021] The method provided by the application takes a reverse perspective from the product to the reactant, first extracts ungenerated products (i.e., products not covered by the initial rule), then traces precursor molecules (second reactants) based on the molecular conversion rule in the refining field, and finally extracts the group structure features thereof to provide a target for supplementing the rule for generating the product. The flow can accurately identify the rule omission point at the product end, is complementary to the identification of the omission from the reactant end, realizes two-way coverage of the rule from the reactant to the product, further improves the integrity of the rule set, and ensures that the molecular level model can predict the reaction path of all actual generated products.

[0022] In a second aspect, the present application provides a reaction rule automatic generation system of a molecular level model, comprising: an acquisition module, configured to acquire actual raw material data and actual product data of a refinery to be modeled, perform data set classification processing, and obtain an actual reactant set, an actual product set and an actual non-reactant set at a molecular level; wherein the actual reactant set comprises molecules that can participate in a refinery chemical reaction, the actual product set comprises molecules generated in an actual production process, and the actual non-reactant set comprises molecules that cannot participate in the refinery chemical reaction; a generation module, configured to construct an initial rule based on a basic reaction type in the field of refining, combine the actual reactant set and the actual product set, and generate an initial reaction rule set; the initial reaction rule set comprises a plurality of reaction rules, and each reaction rule comprises a reactant screening condition and a product generation condition; a first determination module, configured to determine screened reactants from the actual reactant set and generated products from the actual product set, wherein the screened reactants satisfy the reactant screening condition included in the initial reaction rule set, and the generated products satisfy the product generation condition included in the initial reaction rule set; a second determination module, configured to determine a first reactant set according to the screened reactants, the actual reactant set and the actual non-reactant set; wherein the first reactant set comprises a plurality of first reactants and group structure characteristics of each first reactant, and the first reactant is a molecule that belongs to the actual reactant set, does not belong to the screened reactants, and does not belong to the actual non-reactant set; an update module, configured to match each first reactant with a plurality of reaction rules included in a preset reaction rule set according to the group structure characteristics of each first reactant, add the matched reaction rule of each first reactant to the initial reaction rule set, obtain an updated rule set, and the preset reaction rule set is a standard rule set that is pre-constructed and verified by an actual reaction in the field of refining; the second determination module is further configured to determine a second reactant set according to the generated products, the actual product set and the actual reactant set; wherein the second reactant set comprises a plurality of second reactants and group structure characteristics of each second reactant, and the second reactant is a molecule that belongs to the actual reactant set and generates a product that belongs to the actual product set but does not belong to the generated products; and the update module is further configured to match each second reactant with the plurality of reaction rules included in the preset reaction rule set according to the group structure characteristics of the second reactant, add the matched reaction rule of each second reactant to the updated rule set, and obtain a target rule set until the first reactant set and the second reactant set are empty.

[0023] In a third aspect, an electronic device is provided, the electronic device comprising a memory, one or more processors; the memory coupled with the processors; wherein the memory stores computer program codes, the computer program codes comprising computer instructions that, when executed by the processors, cause the electronic device to perform the method according to any implementation manner of the first aspect.

[0024] In a fourth aspect, a computer-readable storage medium is provided, comprising computer instructions that, when executed on an electronic device, cause the electronic device to perform the method according to any implementation manner of the first aspect.

[0025] In a fifth aspect, a computer program product is provided, which, when executed on a computer, causes the computer to perform the method according to any implementation manner of the first aspect.

[0026] It can be understood that the beneficial effects achieved by the system of the second aspect, the electronic device of the third aspect, the computer-readable storage medium of the fourth aspect, and the computer program product of the fifth aspect can refer to the beneficial effects of the first aspect and any possible design manner thereof, which will not be described here. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1 FIG. 1 is a structural schematic diagram of an electronic device according to an embodiment of the present application; Figure 2 FIG. 2 is a flowchart of a reaction rule automatic generation method of a molecular level model according to an embodiment of the present application; Figure 3 FIG. 3 is a structural schematic diagram of a reaction rule automatic generation system according to an embodiment of the present application. DETAILED DESCRIPTION

[0028] The technical solutions in the embodiments of the present application will be described below with reference to the accompanying drawings. In the description of the present application, unless otherwise specified, “ / ” represents an “or” relationship between the objects before and after it, for example, A / B can represent A or B; “or” in the present application is only a description of the relationship between the objects, which means that there can be three relationships, for example, A or B, which means that A exists alone, A and B exist together, and B exists alone, where A and B can be singular or plural. In addition, in the description of the present application, unless otherwise specified, “multiple” means two or more than two. “At least one (one)” or the like means any combination of these items, including any combination of single (one) or multiple items.

[0029] In addition, in order to clearly describe the technical solutions of the embodiments of the present application, in the embodiments of the present application, the terms "first", "second", etc. are used to distinguish the same or similar items with basically the same function and role. Those skilled in the art can understand that the terms "first", "second", etc. do not limit the quantity and execution order, and the terms "first", "second", etc. also do not necessarily mean different.

[0030] Meanwhile, in the embodiments of the present application, the words such as "exemplary" or "for example" are used to represent as an example, illustration or description. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as more excellent or more advantageous than other embodiments or design schemes. Rather, the words such as "exemplary" or "for example" are intended to present the relevant concept in a specific manner for understanding.

[0031] Process flow simulation is a mathematical expression of the chemical thermodynamics and kinetics of the actual refinery conversion process by using computer technology, aiming to reproduce the component movement and conversion law through the model, and then to diagnose, predict and guide the actual operation of the device. With the increasing demand for fine production, traditional virtual component or lumped component modeling has been difficult to meet the accuracy requirements, and molecular level modeling has become an inevitable trend.

[0032] CN111899812A develops a product simulation method for a petroleum processing device, which models the process flow to the molecular level. The model is constructed in four steps: feedstock input, reaction rule setting, reactor configuration, and kinetic parameter parameterization. The setting of the reaction rule needs to reflect the actual chemical reaction process, and its setting directly affects the model output effect, which is a key step in molecular level modeling. CN115841852B provides a method and device for determining the reaction rule of catalytic reforming, which describes a method for determining the reaction rule of a catalytic reforming device, and selects and determines the reactants. However, it does not elaborate on the reactions that actually occur but may be omitted from the rule set, and does not detail the product generation rule, which requires a high level of refinery chemical background for modeling personnel. CN115831248A also provides a method and device for determining the reaction rule, an electronic device, and a storage medium, and CN115841851A provides a method and device for constructing the reaction rule of molecular level hydrocracking.

[0033] However, in the prior art, the supplement of the reaction rule relies on manual judgment by professionals. The missing reactions in the model (such as unreacted reactants and ungenerated products) need to be identified based on experience, and the rules need to be added manually. This not only has low efficiency, but also requires a high level of refinery chemical background for modeling personnel, making it difficult for non-professionals to operate and limiting the popularization and application of molecular level models. In addition, manual rule supplementing is prone to omission or misjudgment, affecting the stability of model accuracy.

[0034] Therefore, there is an urgent need for a reaction rule generation method capable of automatically identifying missing reaction rules without the intervention of professionals to reduce the modeling threshold and improve modeling efficiency and accuracy.

[0035] In view of this, the embodiments of the present application provide a reaction rule automatic generation method and system of a molecular level model, the method comprising: obtaining actual raw material data and actual product data of a to-be-modeled oil refining device, performing data set classification processing to obtain actual reactant set, actual product set and actual non-reacting reactant set at the molecular level; wherein the actual reactant set includes molecules that can participate in oil refining chemical reactions, the actual product set includes molecules generated in the actual production process, and the actual non-reacting reactant set includes molecules that cannot participate in oil refining chemical reactions; constructing an initial rule based on the basic reaction type in the oil refining field, combining the actual reactant set and the actual product set to generate an initialization reaction rule set; the initialization reaction rule set includes a plurality of reaction rules, and each reaction rule includes a reactant screening condition and a product generation condition; determining screened reactants from the actual reactant set and determining generated products from the actual product set, wherein the screened reactants meet the reactant screening condition included in the initialization reaction rule set, and the generated products meet the product generation condition included in the initialization reaction rule set; determining a first reactant set according to the screened reactants, the actual reactant set and the actual non-reacting reactant set; wherein the first reactant set contains a plurality of first reactants and the group structure characteristics of each first reactant, and the first reactant is a molecule belonging to the actual reactant set, not belonging to the screened reactants and not belonging to the actual non-reacting reactant set; according to the group structure characteristics of each first reactant, matching with a plurality of reaction rules contained in a preset reaction rule set, adding the matched reaction rule of each first reactant to the initialization reaction rule set to obtain an updated rule set, and the preset reaction rule set is a standardized rule set constructed in advance and verified by actual reactions in the oil refining field; determining a second reactant set according to the generated products, the actual product set and the actual reactant set; wherein the second reactant set contains a plurality of second reactants and the group structure characteristics of each second reactant, and the second reactant is a molecule belonging to the actual reactant set and generating a product belonging to the actual product set but not belonging to the generated products; according to the group structure characteristics of the second reactants, matching with a plurality of reaction rules contained in the preset reaction rule set, adding the matched reaction rule of each second reactant to the updated rule set until the first reactant set and the second reactant set are empty, to obtain a target rule set.

[0036] The method provided by the present application solves the problem of manual construction of reaction rules in traditional molecular level models by designing a whole process including data acquisition and classification, initial rule set construction, post-filtering reactant / product determination, first / second reactant set supplement, rule iteration and update until the reactant set is empty. In the traditional method, manual rule construction is not only inefficient, but also prone to incomplete rule coverage due to the limitations of artificial experience, such as missing some reactive molecules or corresponding rules for actual product generation. However, the method provided by the present application can automatically classify accurate reactant / product sets from actual data, identify reactive molecules not covered by the initial rules through the first reactant set, and trace back the reactants corresponding to the non-generated products through the second reactant set. Finally, the rules are iteratively completed until there is no omission, and the automatic and comprehensive generation of reaction rules is realized. This method not only avoids the subjectivity and limitations of manual construction, but also significantly improves the efficiency and completeness of rule generation, meeting the core needs of precise and comprehensive reaction rules for molecular level models.

[0037] In some embodiments, the reaction rule automatic generation method for a molecular level model provided by the present application can be executed by a reaction rule automatic generation system 100 (hereinafter referred to as reaction rule automatic generation system 100) for a molecular level model.

[0038] As an example, the reaction rule automatic generation system 100 can be any electronic device 200 with data processing capability, such as a general-purpose computer, a personal computer, a notebook computer, a switch, or a tablet computer, etc. The specific implementation of the reaction rule automatic generation system 100 is not limited here.

[0039] Figure 1 The hardware structure schematic diagram of the electronic device provided by the present application is shown. The electronic device 200 includes a processor 210, a memory 220, and a communication interface 230.

[0040] The processor 210 can include one or more processing cores. The processor 210 connects various parts within the electronic device 200 with various interfaces and lines, performs various functions of the electronic device 200 and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 220, and calling data stored in the memory 220. Alternatively, the processor 210 can be implemented in at least one of a hardware form of a central processing unit (CPU), a graphics processing unit (GPU), a digital signal processing (DSP), a field-programmable gate array (FPGA), a programmable logic array (PLA).

[0041] The memory 220 can include a random access memory (RAM) and can also include a read-only memory (ROM). Alternatively, the memory 220 includes a non-transitory computer-readable storage medium. The memory 220 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 220 can include a program storage area. The program storage area can store instructions for implementing an operating system, instructions for implementing at least one function, instructions for implementing each of the above-mentioned method embodiments, and the like.

[0042] The communication interface 230 is configured to communicate with other devices, apparatuses or communication networks, such as data storage devices, image processing apparatuses or Ethernet, a radio access network (RAN), a wireless local area network (WLAN), and the like.

[0043] In physical implementation, each of the above-mentioned devices (such as the processor 210, the memory 220 and the communication interface 230) can be a device in the same device (such as a notebook computer). Alternatively, at least two of the devices can be arranged in the same device as different devices in the device, similar to the deployment of devices or components in a distributed system.

[0044] It can be understood that the structure illustrated in the embodiment does not constitute a specific limitation on the electronic device 200. In other embodiments of the application, the electronic device 200 can include more or fewer components than illustrated, or combine certain components, or split certain components, or different arrangement of components. The illustrated components can be implemented in hardware, software, or a combination of software and hardware.

[0045] The reaction rule automatic generation method of the molecular level model provided by the embodiment of the application is described below in combination with the accompanying drawings of the specification.

[0046] Figure 2 A flowchart of the reaction rule automatic generation method of the molecular level model provided by the embodiment of the application. Optionally, the method can be executed by the electronic device 200 shown in the figure. The method can include the following steps: Figure 1 S1, obtaining actual raw material data and actual product data of a refinery to be modeled, performing data set classification processing to obtain actual reactant set, actual product set and actual non-reactive reactant set at the molecular level.

[0047] The actual reactant set includes molecules that can participate in oil refining chemical reactions, the actual product set includes molecules generated in the actual production process, and the actual non-reactive reactant set includes molecules that cannot participate in oil refining chemical reactions. In a possible implementation, the actual raw material data at least includes structure-oriented lumped characteristics and mass fraction of each molecule in the raw material of the refinery to be modeled, and the actual product data at least includes structure-oriented lumped characteristics and mass fraction of each molecule in the actual production product of the refinery to be modeled; the data set classification processing adopts a structure-oriented lumped classification algorithm, and by identifying whether the structure-oriented lumped characteristics of the molecule contain active groups that can participate in oil refining reactions, the molecules containing active groups in the actual raw material data are divided into the actual reactant set, and the inert molecules not containing active groups are divided into the actual non-reactive reactant set; at the same time, by comparing the structure-oriented lumped characteristics of the molecule with the actual production record, the molecules generated in the actual production process are screened into the actual product set.

[0048] ​It should be noted that the structure-oriented lumping feature (Structure-Oriented Lumping Feature, SOL feature for short) refers to a set of key structural units that can represent the molecular core reaction activity and structural properties in the oil refining field, which specifically includes functional group types, carbon chain length and branching degree, ring structure number and type, heteroatom (S / N / O) substitution position and number, and other structural information that can directly affect the chemical reaction activity. For example: for alkane molecules, the SOL features include carbon chain length (such as C5-C12 straight chain / branched chain), branched chain position (such as methyl branched chain, ethyl branched chain); for cyclic hydrocarbon molecules, including ring type (six-membered ring, five-membered ring), ring number (single ring, double ring), number of alkyl substituents on the ring; for heteroatom-containing molecules, including heteroatom type (S / N / O), functional group containing heteroatom (such as thiophene ring, pyridine ring, hydroxyl group) and the like.

[0049] The method provided by the present application clearly indicates that the actual raw material and product data need to include structure-oriented lumping (SOL) features and mass fraction, and limits the use of structure-oriented lumping classification algorithm for data processing, and classifies the reactants by identifying active groups and screens the products by comparing production records. In the method provided by the present application, the structure-oriented lumping feature can accurately reflect the core reaction structure of the molecule, the mass fraction provides a basis for data quantification, and the structure-oriented lumping classification algorithm ensures that only molecules containing active groups are classified as reactable reactants, and only molecules in actual production records are classified as products, thereby significantly improving the accuracy of data processing, ensuring the authenticity and effectiveness of the actual reactant set and the product set, providing a reliable data basis for subsequent initialization rule construction, screening and matching, and reducing rule deviation caused by inaccurate data.

[0050] It should be noted that in the determination process of the actual reactant set, the reactants included in the actual reactant set have the following characteristics: the reactants are real molecules, that is, the reactants can be found in the actual reactant set. If the molecule is in the molecular database and also meets the screening rules, but cannot be found in the actual reactant, it is the first excluded set (including multiple third reactants).

[0051] Further, the product is a real molecule. The reaction rule set is constructed and initialized, and the product is generated according to the product generation rule in the reaction rule. It should be in the actual product set, if it cannot be found in the actual product set, the corresponding reactant is the second excluded set (including multiple fourth reactants). Extract the common characteristics of the excluded reactants (third reactants and fourth reactants), manually add reaction rules and limit the screening range of reactants. Until the third reactant and the fourth reactant disappear.

[0052] And the reaction rate of the reactant is less than a preset threshold (0.5X10 -8The above, by preset threshold extraction of the to-be-removed reactant (fifth reactant) characteristics, artificially add reaction rules, until the fifth reactant disappears, obtain the actual reactant set.

[0053] S2, based on the basic reaction type in the oil refining field, the actual reactant set and the actual product set are combined to generate an initialization reaction rule set.

[0054] Specifically, the initialization reaction rule set includes a plurality of reaction rules, and each reaction rule includes a reactant screening condition and a product generation condition. In some embodiments, the basic reaction type in the oil refining field includes a six-membered ring dehydrogenation aromatization reaction and a five-membered ring isomerization reaction, the reactant screening condition of the six-membered ring dehydrogenation aromatization reaction is that the molecule contains a six-membered ring structure feature, the reactant screening condition of the five-membered ring isomerization reaction is that the molecule contains a five-membered ring structure feature, the product generation condition of the six-membered ring dehydrogenation aromatization reaction is to generate a molecule containing a monocyclic aromatic hydrocarbon structure feature, and the product generation condition of the five-membered ring isomerization reaction is to generate a molecule containing a five-membered ring derivative structure feature.

[0055] The method provided by the application clearly defines the basic reaction type in the oil refining field as a six-membered ring dehydrogenation aromatization reaction and a five-membered ring isomerization reaction, and directly corresponds to the reactant screening conditions and product generation conditions of the two types of reactions. The method provided by the application focuses on the core ring reaction in the oil refining field, and directly binds the reaction type and the molecule SOL feature, so that the screening conditions and generation conditions of the initial rules are visualized and can be implemented on the ground, avoiding the ambiguity of the initial rules, providing clear and executable basis for the determination of the screened reactants or products, and ensuring that the initial rules can accurately cover the key basic reactions in the oil refining field, laying a clear starting point for rule iteration.

[0056] S3, determining a screened reactant from the actual reactant set and a generated product from the actual product set.

[0057] The screened reactant satisfies the reactant screening condition included in the initialization reaction rule set, and the generated product satisfies the product generation condition included in the initialization reaction rule set.

[0058] In one possible implementation, the determination of the screened reactant from the actual reactant set includes: The preset reactant screening program is called, the reactant screening program loads the reactant screening condition corresponding to each reaction rule in the initialization reaction rule set, and the structure-oriented collective feature matching is performed on each molecule in the actual reactant set, and the molecules that pass the matching are determined as the screened reactants; In another possible implementation, the determining the generated post-product from the actual product set comprises: calling a preset product generation program, the product generation program simulating a molecular conversion process of the screened reactant based on a product generation condition corresponding to each reaction rule in the initialized reaction rule set to generate a simulation product, and then performing a structure-oriented collective feature comparison between the simulation product and the actual product set to determine the simulation product existing in the actual product set as the generated post-product.

[0059] In the method provided by the application, the preset program can quickly and accurately screen the reactant conforming to the rule through automatic matching of the SOL feature, the product generation program can ensure that the generated post-product is a molecule existing in actual production through double verification of simulation conversion and actual comparison, which not only reduces errors caused by manual intervention, but also greatly improves the efficiency of the screening and generation process, and provides reliable intermediate results for subsequent identification of the first / second reactant set, thereby avoiding deviation of the entire rule generation process caused by errors in the intermediate results.

[0060] S4, determining the first reactant set according to the screened reactant, the actual reactant set and the actual non-reactive reactant set.

[0061] The first reactant set includes a plurality of first reactants and group structure features of each first reactant, and the first reactant is a molecule belonging to the actual reactant set, not belonging to the screened reactant and not belonging to the actual non-reactive reactant set. It should be understood that the first reactant is the unreacted reactant, that is, such a reactant has reacted in actual production, but due to the lack of rules or too tight definition of corresponding rules, it does not react in the existing rule model.

[0062] In some embodiments, the above S5 specifically comprises: comparing the screened reactant with the actual reactant set to extract molecules in the actual reactant set that are not included in the screened reactant to obtain an unselected molecule list; comparing the unselected molecule list with the actual non-reactive reactant set to eliminate inert molecules in the unselected molecule list belonging to the actual non-reactive reactant set, and the remaining molecules constitute the first reactant; performing structure-oriented collective feature extraction on each first reactant to obtain group structure features of each first reactant, and determining the first reactant set according to the group structure features of each first reactant.

[0063] In the method provided by the application, the first step accurately locates the molecules not covered by the initial rules by comparing and screening the actual reactants with the reactant set; the second step removes the non-reactive molecules to ensure that the remaining molecules are all effective molecules that can react but are not covered; and the third step extracts the group structure characteristics to provide clear targeting information for subsequent rule matching and rule supplement. The process can accurately identify the missing points of the initial rules, avoid the lack of rules caused by the non-covered reactive molecules, and provide clear basis for rule completion to ensure the pertinence of subsequent rule iteration.

[0064] S5, according to the group structure characteristics of each of the first reactants, matching a plurality of reaction rules contained in the preset reaction rule set, adding the reaction rule matched by each of the first reactants to the initial reaction rule set to obtain an updated rule set.

[0065] The preset reaction rule set is a standardized rule set constructed in advance and verified by actual reactions in the oil refining field. In some embodiments, the preset reaction rule set includes alkyl isomerization reaction rules, alkyl dehydrogenation cyclization reaction rules, alkyl hydrocracking reaction rules, five-membered ring isomerization reaction rules, six-membered ring dehydrogenation aromatization reaction rules, hydrodesulfurization reaction rules, hydrodenitrogenation reaction rules, and hydrodeoxygenation reaction rules. Each type of reaction rule contains group matching conditions and molecular conversion paths, wherein the group matching conditions are constraints corresponding to the group structure characteristics of the first reactants, and the molecular conversion paths are standard oil refining chemical reaction paths corresponding to the groups.

[0066] The eight types of rules covered by the method provided by the application cover the main reaction types in the oil refining field, including alkyl reactions, ring reactions, and heteroatom removal, ensuring comprehensive rule resources. The group matching conditions provide clear standards for the matching of the first reactants and the rules, and the molecular conversion paths ensure that the rules comply with the oil refining reaction rules, thereby avoiding the randomness of rule matching and improving the effectiveness of rule supplement, providing reliable and standardized resource support for the rule completion of the first reactants.

[0067] In one possible implementation, the above S5 includes: analyzing the group structure characteristics of each first reactant to determine the core reaction group of each first reactant; in the preset reaction rule set, searching for a reaction rule whose group matching condition completely corresponds to the core reaction group; if the corresponding reaction rule is found and does not exist in the initial reaction rule set, the corresponding reaction rule is added to the initial reaction rule set to complete the rule supplement and form an updated rule set.

[0068] The method provided by the application is characterized in that the rule for ensuring matching of the parsed core reaction group is for the key reaction site of the first reactant, the rule is ensured to meet the preset standard, and the rule is added only when the rule does not exist, so that the rule set is kept simple. The process realizes accurate and non-redundant supplement of the rule, ensures that the supplemented rule can accurately cover the reaction corresponding to the first reactant, avoids the subsequent efficiency reduction caused by the bloated rule set, and improves the pertinence and practicality of the updated rule set.

[0069] It should be noted that the application screens, classifies and extracts the structural unit features of the first reactant by a self-set program, relaxes or supplements new reaction rules to the existing rules by calling the preset reaction rule set and program comparison until the first reactant disappears, and realizes automatic updating of the existing reaction rules and automatic supplement of the missing reaction rules.

[0070] S6, determining a second reactant set according to the generated product, the actual product set and the actual reactant set.

[0071] The second reactant set includes a plurality of second reactants and group structure features of each second reactant, and the second reactant belongs to the actual reactant set and generates a product that belongs to the actual product set but does not belong to the generated product. It should be understood that the second reactant is a product that is not generated, that is, such a product is generated in practice, and the reactant corresponding to the product has been screened out to correspond to a certain rule, but the rule is too tight for product generation, so that the product is not generated in the existing rule model.

[0072] In a possible implementation, the determining of the second reactant set according to the generated product, the actual product set and the actual reactant set includes: comparing the generated product with the actual product set, extracting molecules in the actual product set that are not included in the generated product to obtain a list of ungenerated products; based on the molecular conversion rule in the oil refining field, tracing precursor molecules that can generate each molecule in the list of ungenerated products in the actual reactant set; performing structure-oriented aggregate feature extraction on each precursor molecule to obtain group structure features of each precursor molecule, and determining the second reactant set according to the group structure features of each precursor molecule.

[0073] Specifically, molecular transformation rules in the oil refining field refer to the set of molecular structure transformation rules that have been experimentally verified and widely recognized during the oil refining process. These rules include three core categories: functional group transformation rules, carbon chain rearrangement rules, and heteroatom removal rules. These rules serve as the basis for tracing precursor molecules that have not yet formed products. The specific content is as follows: For functional group transformation rules, this includes ring structure transformation: six-membered cycloalkanes can be transformed into monocyclic aromatic hydrocarbons through dehydrogenation reactions; five-membered cycloalkanes can be transformed into six-membered cycloalkanes through isomerization reactions; monocyclic aromatic hydrocarbons can be transformed into six-membered cycloalkanes through hydrogenation reactions; and unsaturated bond transformation: alkenes can be transformed into alkanes through hydrogenation reactions; and alkynes can be transformed into alkenes through partial hydrogenation reactions. Hydrocarbons; heteroatom functional group transformation: Thiophene (containing S) can be converted into alkanes and H2S through hydrodesulfurization; pyridine (containing N) can be converted into alkanes and NH3 through hydrodenitrogenation; phenols (containing O) can be converted into alkanes and H2O through hydrodeoxygenation. For carbon chain rearrangement rules, these include: alkyl branched rearrangement: straight-chain alkanes can be converted into branched-chain alkanes through isomerization; the branch positions of branched alkanes can migrate; carbon chain cleavage: long-chain alkanes (C≥10) can be cleaved into short-chain alkanes through hydrocracking; cycloalkanes can have their side chains cleaved into independent alkane molecules. For heteroatom removal rules, these include: desulfurization rules, denitrogenation rules, and deoxygenation rules.

[0074] The method provided by this invention adopts a reverse perspective from product to reactant. It first extracts unformed products (i.e., products not covered by the initial rules), then traces precursor molecules (second reactants) based on the molecular transformation rules of oil refining, and finally extracts their functional group structural features, providing a target for supplementing rules that generate these products. This process can accurately identify rule omissions at the product end, complementing the identification of omissions at the reactant end, achieving bidirectional rule coverage from reactant to product, further improving the completeness of the rule set, and ensuring that the molecular-level model can predict the reaction paths of all actual products. This invention uses a self-designed program to screen, classify, and extract the structural unit features of second reactants. By calling a preset reaction rule set and comparing it with the program, it relaxes existing rules or supplements them with new reaction rules until all second reactants disappear, achieving automatic updating of existing reaction rules and automatic supplementation of omitted reaction rules.

[0075] S7. Based on the group structure characteristics of the second reactant, match it with multiple reaction rules contained in the preset reaction rule set, and add the reaction rule matched by each second reactant to the updated rule set until both the first reactant set and the second reactant set are empty, thus obtaining the target rule set.

[0076] The application realizes automatic updating of existing reaction rules and automatic supplement of missing reaction rules by setting programs, screening, classifying and extracting structural unit characteristics of the second reactants, calling a preset reaction rule set, and relaxing or supplementing new reaction rules to the existing rules until the second reactants disappear.

[0077] As can be seen from S1-S7, the method provided by the embodiments of the application solves the problem of manual construction of reaction rules in a traditional molecular level model by designing a full process of data acquisition and classification, initial rule set construction, determination of screened reactants / products, supplement of the first / second reactant set, and iterative updating of rules until the reactant set is empty. In the traditional method, manual rule construction is not only inefficient, but also prone to incomplete rule coverage due to the limitations of artificial experience, such as missing some reactionable molecules or corresponding rules for actually generated products. However, the method provided by the application can first obtain accurate reactant / product sets from actual data through an automatic process, then identify reactionable molecules that are not covered by the initial rules through the first reactant set, and finally iteratively complete the rules until there is no missing rule, so as to realize automatic and comprehensive generation of reaction rules, avoid subjectivity and limitations of manual construction, significantly improve the efficiency and completeness of rule generation, and meet the core needs of the molecular level model for accurate and comprehensive reaction rules.

[0078] It can also be understood that the method provided by the application is applied to automatic generation of missing reaction rules in the modeling process of a molecular level model of a refining and chemical device. By automatically supplementing the missing reaction rules, the method ensures that the reactant screening rules are not missed, and all actually occurring reactions are automatically included in the reaction rules of the molecular level modeling, which helps professional or non-professional personnel quickly construct reaction rules, speeds up the modeling process, and lays a technical foundation for improving the model accuracy.

[0079] In order to facilitate understanding of the present scheme, the method proposed by the embodiments of the application will be explained and described below in combination with an example.

[0080] In one example, the above method includes: Based on the actual raw materials and products of the refining device to be modeled, a molecular level actual reactant set, an actual product set, and an actual non-reactive reactant set are constructed.

[0081] A reactant screening-product generation rule set (i.e., an initial reaction rule set) is constructed and initialized.

[0082] A reactant screening program is called to complete reactant screening: this is a program for automatically screening reactants according to certain rules, which is self-contained by the system, and its function is to ensure that the screened reactants are real molecules in the actual reactant set.

[0083] Call product generation program to complete product generation determination: this is a program that automatically determines product generation rules according to certain rules, which is a system itself, and its function is to ensure that the screened products are real molecules, in the actual product set, and their generation conforms to the actual chemical reaction.

[0084] Perform first reactant determination, specifically including: Automatically extract unselected reactants in the actual reactant set. Through the self-set program, extract the unreacted reactants that are not selected by the current rule. Through the self-set program and the actual unreactant set comparison, extract the first reactants (unreacted reactants) that are truly occurring but not reacted by the current rule, to form the first reactant set. Extract the molecular group structure characteristics in the first reactant set.

[0085] Call the preset reaction rule set: the reaction rules under this rule set are the molecular level model modeling process accumulated and verified for many times, which includes: alkyl isomerization reaction rule, alkyl dehydrogenation cyclization reaction rule, alkyl hydrogenation cracking reaction rule, five-membered ring isomerization reaction rule, six-membered ring dehydrogenation aromatization reaction rule, hydrogenation desulfurization reaction rule, hydrogenation denitrification reaction rule, hydrogenation deoxygenation reaction rule.

[0086] Supplement reaction rules: search for corresponding reaction rules in the preset reaction rule set according to the group characteristics, if the current rule already exists the same type of rule, then replace the current rule with the corresponding reaction rule in the preset reaction rule set. If the current rule lacks such reaction rules, then introduce and supplement this rule to the current reaction rule.

[0087] Perform second reactant determination, specifically including: Automatically extract unselected reactants in the actual reactant set. Through the self-set program, extract the unreacted reactants that are not selected by the current rule. Through the self-set program and the actual unreactant set comparison, extract the first reactants (unreacted reactants) that are truly occurring but not reacted by the current rule, to form the first reactant set. Extract the molecular group structure characteristics in the first reactant set.

[0088] Automatically loop to call the preset reaction rule set, supplement the reaction rules and perform the second reactant determination operation until all first reactants and second reactants are eliminated.

[0089] Get the target reactant screening-product generation rule set (also known as the target rule set). That is: the reaction rule set has covered all the reactants and all the generated products.

[0090] In another example, taking the actual reactant set = {1, 2, 3, 4, 5}, the actual product set = {6, 7, 8, 9, 10}, and the actual non-reactive reactant set = {11, 12} as an example, the complete execution logic of this scheme is as follows: first, step S3 is executed for the first screening and generation, based on the initial constructed initialization reaction rule set, only the screened reactants = {1, 4} that meet the reactant screening conditions can be screened out, and the generated products = {6, 9} that meet the product generation conditions and exist in the actual product set are simulated; then step S4 is entered, by comparing the screened reactants with the actual reactant set, the molecules {2, 3, 5} that are not screened are extracted, and the inert molecules in the actual non-reactive reactant set {11, 12} are removed, and finally the first reactant set = {2, 3, 5} is determined, which are all valid molecules belonging to the actual reactant set and can participate in the reaction but are not covered by the initial rules; then in step S5, based on the group structure characteristics of each molecule in the first reactant set, the preset reaction rule set is matched, and rules 3 (alkyl isomerization), 4 (hydrogenation desulfurization), and 5 (hydrogenation denitrogenation) are supplemented to the initialization reaction rule set to obtain the updated rule set; after entering step S6, the generated products = {6, 9} are compared with the actual product set = {6, 7, 8, 9, 10}, the non-generated products = {7, 8, 10} are extracted, and the precursor molecules that can generate these products are still {2, 3, 5}, so the second reactant set = {2, 3, 5} is determined, which are all molecules belonging to the actual reactant set and generating products belonging to the actual product set but not included in the first generated products; finally, step S7 is entered into the loop iteration process, after the supplemented rules are added to the updated rule set, the screening and generation are re-executed in step S3, at this time the updated rule set can cover all actual reactants and products, the screened reactants = {1, 2, 3, 4, 5} and the generated products = {6, 7, 8, 9, 10} are obtained, and step S4 is executed again, because the actual reactant set has been included in the screened reactants, the first reactant set is empty; when step S6 is executed, because the actual product set has been included in the generated products, the second reactant set is empty, which meets the iteration termination condition, and finally the target rule set covering all reactants and products is output, and the whole process logic is completely closed loop.

[0091] The above describes the scheme of the embodiments of the present application mainly from the perspective of the method. It can be understood that the reaction rule automatic generation system 100 comprises at least one of the hardware structure and the software module corresponding to each function in order to realize the above functions. Those skilled in the art should easily realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized in the form of hardware or the combination of hardware and computer software. Whether a certain function is realized in the form of hardware or computer software driven hardware depends on the specific application and design constraints of the technical scheme. The skilled person can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the embodiments of the present application.

[0092] The embodiments of the present application can divide the reaction rule automatic generation system 100 into functional units according to the above method examples. For example, the reaction rule automatic generation system 100 can be divided into functional units corresponding to each function, or two or more functions can be integrated into one processing unit. The above integrated unit can be realized in the form of hardware or software functional unit. It should be noted that the division of units in the embodiments of the present application is illustrative, and is only a logical functional division. There can be another division method in actual implementation.

[0093] For example, Figure 3A hardware structure schematic diagram of a reaction rule automatic generation system provided by an embodiment of the present application is shown. The reaction rule automatic generation system 100 comprises: an acquisition module 110, configured to acquire actual raw material data and actual product data of a refinery to be modeled, perform data set classification processing, and obtain actual reactant sets, actual product sets and actual non-reacting reactant sets at a molecular level; wherein the actual reactant sets comprise molecules that can participate in a refinery chemical reaction, the actual product sets comprise molecules generated in an actual production process, and the actual non-reacting reactant sets comprise molecules that cannot participate in a refinery chemical reaction; a generation module 120, configured to construct an initial rule based on a basic reaction type in the field of refining, combine the actual reactant sets and the actual product sets, and generate an initial reaction rule set; the initial reaction rule set comprises a plurality of reaction rules, and each reaction rule comprises a reactant screening condition and a product generation condition; a first determination module 130, configured to determine screened reactants from the actual reactant sets and generated products from the actual product sets, wherein the screened reactants satisfy the reactant screening condition included in the initial reaction rule set, and the generated products satisfy the product generation condition included in the initial reaction rule set; a second determination module 140, configured to determine a first reactant set according to the screened reactants, the actual reactant sets and the actual non-reacting reactant sets; wherein the first reactant set comprises a plurality of first reactants and group structure characteristics of each first reactant, and the first reactant is a molecule that belongs to the actual reactant sets, does not belong to the screened reactants and does not belong to the actual non-reacting reactant sets; an update module 150, configured to match each first reactant with a plurality of reaction rules included in a preset reaction rule set according to the group structure characteristics of each first reactant, add the reaction rule matched by each first reactant to the initial reaction rule set, obtain an updated rule set, and the preset reaction rule set is a standard rule set that is constructed in advance and verified by actual reactions in the field of refining; the second determination module 140 is further configured to determine a second reactant set according to the generated products, the actual product sets and the actual reactant sets; wherein the second reactant set comprises a plurality of second reactants and group structure characteristics of each second reactant, and the second reactant is a molecule that belongs to the actual reactant sets and generates a product that belongs to the actual product sets but does not belong to the generated products; and the update module 150 is further configured to match each second reactant with the plurality of reaction rules included in the preset reaction rule set according to the group structure characteristics of the second reactant, add the reaction rule matched by each second reactant to the updated rule set, and obtain a target rule set until the first reactant set and the second reactant set are empty.

[0094] It should be understood that the specific description of the above optional mode can refer to the foregoing method embodiments, and will not be described here. In addition, the description of the above explanation and beneficial effects of any one of the reaction rule automatic generation system 100 provided can refer to the corresponding method embodiments described above, and will not be described here.

[0095] The embodiments of the present application also provide a computer readable storage medium, which stores at least one computer instruction. The at least one computer instruction is loaded and executed by a processor to implement the method of each of the above embodiments. The description of the explanation and beneficial effects of any one of the computer readable storage medium provided above can refer to the corresponding embodiments described above, and will not be described here.

[0096] The embodiments of the present application also provide a chip. The chip integrates a control circuit and one or more ports for implementing the functions of the above reaction rule automatic generation system 100. Optionally, the functions supported by the chip can refer to the above, and will not be described here.

[0097] Those skilled in the art can understand that all or part of the steps of the above embodiments can be completed by a program instructing related hardware. The program can be stored in a computer readable storage medium. The storage medium mentioned above can be a read-only memory, a random access memory, etc. The processing unit or the processor can be a central processing unit, a general-purpose processor, an application specific integrated circuit (ASIC), a microprocessor (digital signal processor, DSP), a field programmable gate array (FPGA) or other programmable logic device, a transistor logic device, a hardware component or any combination thereof.

[0098] The embodiments of the present application also provide a computer program product containing instructions, which, when executed on a computer, cause the computer to perform any of the methods described above. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on the computer, the processes or functions according to the embodiments of the present application are wholly or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable apparatus. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center through a wired (for example, coaxial cable, optical fiber, digital subscriber line (DSL), or wireless (for example, infrared, wireless, microwave, etc.) manner. The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server, data center, etc. containing one or more available media sets. The available media can be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium (for example, an SSD), etc.

[0099] It should be noted that the above-mentioned devices for storing computer instructions or computer programs provided by the embodiments of the present application, such as but not limited to the above-mentioned memory, computer-readable storage medium, and communication chip, etc., are all non-transitory. Those skilled in the art should be aware that in one or more of the above examples, the functions described by the embodiments of the present application can be implemented by hardware, software, firmware, or any combination thereof. When implemented by software, these functions can be stored in a computer-readable storage medium or transmitted as one or more instructions or codes on a computer-readable storage medium. The computer-readable storage medium includes computer storage medium and communication medium, wherein the communication medium includes any medium that facilitates the transfer of a computer program from one place to another. The storage medium can be any available medium accessible by a general-purpose or special-purpose computer.

[0100] Although the embodiments of the present application have been shown and described above, it should be understood that the above-mentioned embodiments are exemplary and should not be construed as limiting the present application, and those skilled in the art can make changes, modifications, replacements, and variations to the above-mentioned embodiments within the scope of the present application.

Claims

1. A method for automatically generating reaction rules in a molecular-level model, characterized in that, The method includes: The actual feedstock data and actual product data of the oil refining unit to be modeled are obtained, and the dataset is classified and processed to obtain the molecular-level actual reactant set, actual product set, and actual non-reactant set; wherein, the actual reactant set includes molecules that can participate in oil refining chemical reactions, the actual product set includes molecules generated in the actual production process, and the actual non-reactant set includes molecules that cannot participate in oil refining chemical reactions. Initial rules are constructed based on the basic reaction types in the oil refining field. Combined with the actual set of reactants and the actual set of products, an initial reaction rule set is generated. The initial reaction rule set includes multiple reaction rules, and each reaction rule includes reactant screening conditions and product generation conditions. Screened reactants are determined from the actual reactant set, and generated products are determined from the actual product set, wherein the screened reactants satisfy the reactant screening conditions included in the initial reaction rule set, and the generated products satisfy the product generation conditions included in the initial reaction rule set. Based on the screened reactants, the actual reactant set, and the actual non-reactant set, a first reactant set is determined; wherein, the first reactant set includes multiple first reactants and the group structure features of each first reactant, and the first reactants are molecules that belong to the actual reactant set, but do not belong to the screened reactants and do not belong to the actual non-reactant set; Based on the group structure characteristics of each first reactant, it is matched with multiple reaction rules contained in the preset reaction rule set. The reaction rule matched by each first reactant is added to the initial reaction rule set to obtain the updated rule set. The preset reaction rule set is a standardized rule set that has been pre-constructed and verified by actual reactions in the oil refining field. Based on the generated product, the actual product set, and the actual reactant set, a second reactant set is determined; wherein, the second reactant set includes multiple second reactants and the group structure features of each second reactant, and the second reactant is a molecule that belongs to the actual reactant set and whose generated product belongs to the actual product set but not to the generated product; Based on the functional group structure characteristics of the second reactant, it is matched with multiple reaction rules contained in the preset reaction rule set. Each reaction rule matched by the second reactant is added to the updated rule set until both the first reactant set and the second reactant set are empty, thus obtaining the target rule set.

2. The method according to claim 1, characterized in that, The actual raw material data includes at least the structure-oriented aggregate features and mass fractions of each molecule in the raw material of the refinery unit to be modeled, and the actual product data includes at least the structure-oriented aggregate features and mass fractions of each molecule in the actual production products of the refinery unit to be modeled. The dataset classification process adopts a structure-oriented aggregate classification algorithm, which identifies whether the structure-oriented aggregate features of molecules contain active groups that can participate in the refining reaction. Molecules containing active groups in the actual raw material data are classified into the actual reactant set, and inert molecules without active groups are classified into the actual non-reactant set. At the same time, by comparing the structure-oriented aggregate features of molecules with the actual production records, molecules generated in the actual production process are screened into the actual product set.

3. The method according to claim 2, characterized in that, The basic reaction types in the oil refining field include six-membered ring dehydrogenation aromatization and five-membered ring isomerization. The reactant selection condition for the six-membered ring dehydrogenation aromatization reaction is that the molecule contains a six-membered ring structure. The reactant selection condition for the five-membered ring isomerization reaction is that the molecule contains a five-membered ring structure. The product formation condition for the six-membered ring dehydrogenation aromatization reaction is to generate a molecule containing a monocyclic aromatic hydrocarbon structure. The product formation condition for the five-membered ring isomerization reaction is to generate a molecule containing a five-membered ring derivative structure.

4. The method according to claim 3, characterized in that, The determination of screened reactants from the actual reactant set includes: The preset reactant screening program is invoked. The reactant screening program loads and initializes the reactant screening conditions corresponding to each reaction rule in the reaction rule set, performs structure-guided lumped feature matching on each molecule in the actual reactant set, and determines the successfully matched molecules as the screened reactants. The determination of the generated product from the actual product set includes: A preset product generation program is invoked. The product generation program simulates the molecular transformation process of the screened reactants based on the product generation conditions corresponding to each reaction rule in the initial reaction rule set, generates simulated products, and then compares the simulated products with the actual product set using structure-guided lumped features. The simulated products that exist in the actual product set are identified as the generated products.

5. The method according to claim 4, characterized in that, The step of determining the first reactant set based on the screened reactants, the actual reactant set, and the actual non-reactant set includes: The screened reactants are compared with the actual reactant set, and molecules that were not included in the screened reactants in the actual reactant set are extracted to obtain a list of unscreened molecules. The list of unscreened molecules is compared with the actual set of non-reacting reactants. Inert molecules that belong to the actual set of non-reacting reactants are removed from the list of unscreened molecules, and the remaining molecules constitute the first reactant. For each first reactant, perform structure-guided lumped feature extraction to obtain the group structure features of each first reactant, and determine the first reactant set based on the group structure features of each first reactant.

6. The method according to claim 5, characterized in that, The preset reaction rule set includes alkyl isomerization reaction rules, alkyl dehydrogenation cyclization reaction rules, alkyl hydrocracking reaction rules, five-membered ring isomerization reaction rules, six-membered ring dehydrogenation aromatization reaction rules, hydrodesulfurization reaction rules, hydrodenitrogenation reaction rules, and hydrodeoxygenation reaction rules. Each type of reaction rule includes group matching conditions and molecular transformation paths. The group matching conditions are constraints corresponding to the group structure characteristics of the first reactant, and the molecular transformation paths are the standard refining chemical reaction paths corresponding to that type of group.

7. The method according to claim 6, characterized in that, The step involves matching each first reactant with multiple reaction rules contained in a preset reaction rule set based on the group structure characteristics of each first reactant, and adding the reaction rule matched by each first reactant to the initial reaction rule set to obtain an updated rule set, including: The functional group structure characteristics of each first reactant are analyzed to determine the core reactive group of each first reactant; in the preset reaction rule set, the reaction rule whose functional group matching condition completely corresponds to the core reactive group is searched. If a corresponding reaction rule is found, and the corresponding reaction rule does not exist in the initial reaction rule set, then the corresponding reaction rule is added to the initial reaction rule set to complete the rule supplementation and form an updated rule set.

8. The method according to claim 7, characterized in that, The step of determining the second reactant set based on the generated products, the actual product set, and the actual reactant set includes: The generated products are compared with the actual product set, and molecules that were not included in the generated products are extracted from the actual product set to obtain a list of ungenerated products. Based on the molecular transformation laws in the oil refining field, the precursor molecules of each molecule in the list of molecules that can generate but do not generate products are traced in the actual reactant set. For each precursor molecule, structure-guided lumped feature extraction is performed to obtain the group structure features of each precursor molecule, and the second reactant set is determined based on the group structure features of each precursor molecule.

9. An automatic reaction rule generation system for molecular-level models, characterized in that, The system includes: The acquisition module is used to acquire the actual raw material data and actual product data of the oil refining unit to be modeled, perform dataset classification processing, and obtain the molecular-level actual reactant set, actual product set, and actual non-reactant set; wherein, the actual reactant set includes molecules that can participate in oil refining chemical reactions, the actual product set includes molecules generated in the actual production process, and the actual non-reactant set includes molecules that cannot participate in oil refining chemical reactions. The generation module is used to construct initial rules based on the basic reaction types in the oil refining field, and combine the actual reactant set and the actual product set to generate an initial reaction rule set; the initial reaction rule set includes multiple reaction rules, and each reaction rule includes reactant screening conditions and product generation conditions. The first determining module is used to determine screened reactants from the actual reactant set and generated products from the actual product set, wherein the screened reactants satisfy the reactant screening conditions included in the initial reaction rule set and the generated products satisfy the product generation conditions included in the initial reaction rule set. The second determining module is used to determine a first reactant set based on the screened reactants, the actual reactant set, and the actual non-reactant set; wherein the first reactant set includes multiple first reactants and the group structure features of each first reactant, and the first reactant is a molecule that belongs to the actual reactant set, but does not belong to the screened reactants and does not belong to the actual non-reactant set; The update module is used to match each of the first reactants with multiple reaction rules contained in the preset reaction rule set based on the group structure characteristics of each first reactant, and add the reaction rule matched by each first reactant to the initial reaction rule set to obtain the updated rule set. The preset reaction rule set is a standardized rule set that has been pre-constructed and verified by actual reactions in the oil refining field. The second determining module is further configured to determine a second reactant set based on the generated product, the actual product set, and the actual reactant set; wherein the second reactant set includes multiple second reactants and the group structure features of each second reactant, and the second reactant is a molecule that belongs to the actual reactant set and the generated product belongs to the actual product set but not to the generated product; The updating module is further configured to match the group structure characteristics of the second reactant with multiple reaction rules contained in the preset reaction rule set, and add the reaction rule matched by each second reactant to the updated rule set until both the first reactant set and the second reactant set are empty, thereby obtaining the target rule set.

10. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the automatic generation method of reaction rules for molecular-level models as described in any one of claims 1-8.

Images