CN111899812A

CN111899812A - Product simulation method, system, equipment and storage medium of petroleum processing device

Info

Publication number: CN111899812A
Application number: CN202010533482.2A
Authority: CN
Inventors: 纪晔; 韩景宽; 王杭州; 杨诗棋; 孙宝文; 熊纯青; 李凤琪
Original assignee: Petrochina Co Ltd
Current assignee: Petrochina Co Ltd
Priority date: 2020-06-12
Filing date: 2020-06-12
Publication date: 2020-11-06
Anticipated expiration: 2040-06-12
Also published as: CN111899812B

Abstract

The invention discloses a product simulation method, a product simulation system, a product simulation device and a storage medium of a petroleum processing device. The method comprises the following steps: determining the molecular composition of the feedstock; inputting the molecular composition of the raw material into a pre-trained product prediction model; wherein the product prediction model corresponds to a type of petroleum processing plant; and obtaining the yield of the target product under different reaction conditions output by the product prediction model. The invention correspondingly trains a product prediction model for each petroleum processing device in advance, and predicts the yield of the target product of the corresponding petroleum processing device under different reaction conditions through the corresponding product prediction model. The method has the advantages of high accuracy, short time and low cost.

Description

Product simulation method, system, equipment and storage medium of petroleum processing device

Technical Field

The invention relates to the technical field of computers, in particular to a product simulation method, a product simulation system, a product simulation device and a storage medium of a petroleum processing device.

Background

At present, shortage of petroleum resources is increasingly remarkable, and meanwhile, with the increase of the demand of various countries on clean oil products and the improvement of environmental awareness, the production of clean environment-friendly fuels and the reduction of environmental pollution hazards become urgent. Therefore, researchers have focused on improving existing gasoline production processes or developing new technologies to produce cleaner petroleum products and to maximize the production of aromatics, ethylene, and other important chemical products.

However, the types of petroleum processing devices are various, and when the types of complex and varied raw materials are faced, the reaction process of the raw materials in each petroleum processing device is generally difficult to determine, and the product information of each petroleum processing device is further impossible to determine, so that the components of the products cannot be determined in advance, and the value of the products is difficult to estimate. For example: catalytic reforming is an important naphtha processing technology in the oil refining industry and is also an important process for improving the octane number of gasoline and realizing the integrated goal of oil refining/chemical industry. The catalytic reforming device produces low-sulfur low-olefin clean gasoline and can also produce high-purity hydrogen as a byproduct. The products produced by the catalytic reforming process can meet the demands of the market for clean fuels and aromatic hydrocarbons to a great extent. Therefore, there is a trend in the petrochemical industry to devote catalytic reforming processes. However, in the face of complicated and variable raw materials, it is difficult to accurately reflect the reaction mechanism of the catalytic reforming process, and the product after catalytic reforming cannot be determined.

Disclosure of Invention

The invention mainly aims to provide a product simulation method, a product simulation system, a product simulation device and a storage medium of a petroleum processing device, so as to solve the problem that the product information of the petroleum processing device cannot be determined in the prior art.

Aiming at the technical problems, the invention solves the technical problems by the following technical scheme:

the invention provides a product simulation method of a petroleum processing device, which comprises the following steps: determining the molecular composition of the feedstock; inputting the molecular composition of the raw material into a pre-trained product prediction model; wherein the product prediction model corresponds to a type of petroleum processing plant; and obtaining the yield of the target product under different reaction conditions output by the product prediction model.

After obtaining the yields of the target products under different reaction conditions output by the product prediction model, the method preferably further comprises: inquiring the yield of the target product under different reaction conditions; determining a reaction condition corresponding to the maximum yield value of the target product; and/or determining the reaction condition corresponding to the minimum yield value of the target product.

After obtaining the yields of the target products under different reaction conditions output by the product prediction model, the method preferably further comprises: and determining the physical properties of the target product under different reaction conditions according to the molecular composition of the target product under different reaction conditions.

Preferably, obtaining the yields of the target product under different reaction conditions output by the product prediction model comprises: obtaining the content of each product molecule output by the product prediction model under different reaction conditions; and obtaining the content of the target product molecules under different reaction conditions from the content of each product molecule under different reaction conditions.

Wherein determining the molecular composition of the feedstock preferably comprises: determining the molecular composition of the feedstock by one or more of comprehensive two-dimensional gas chromatography, quadrupole gas chromatography-mass spectrometry detection, gas chromatography/field ionization-time-of-flight mass spectrometry detection, gas chromatography, near infrared spectroscopy, nuclear magnetic resonance spectroscopy, raman spectroscopy, fourier transform ion cyclotron resonance mass spectrometry, electrostatic field orbitrap mass spectrometry, and ion mobility mass spectrometry.

Wherein the step of training the product prediction model preferably comprises: establishing a product prediction model; wherein the product prediction model comprises: a reaction rule set and a reaction rate algorithm; the set of reaction rules comprises a plurality of reaction rules; acquiring sample raw material information of sample raw materials; training the reaction rule set by using the sample raw material information, and fixing the trained reaction rule set; and training the reaction rate algorithm by using the sample raw material information, and fixing the trained reaction rate algorithm to obtain the trained product prediction model.

Wherein, the sample material information of the sample material preferably includes: the molecular composition of the sample raw material, the molecular content of each molecule in the sample raw material, the molecular composition of an actual product corresponding to the sample raw material and the actual content of each molecule in the actual product.

Wherein, using the sample material information to train the reaction rule set, preferably comprising: processing the molecular composition of the sample raw material according to a preset reaction rule set to obtain a reaction path corresponding to each molecule in the molecular composition of the sample raw material; obtaining a first molecular composition of a device product according to a reaction path corresponding to each molecule in the molecular composition of the sample raw material; in the device product, comprising: the sample feedstock, intermediate product, and predicted product; calculating a first relative deviation from a first molecular composition of the device product and a second molecular composition of the actual product; if the first relative deviation meets a preset condition, fixing the reaction rule set; and if the first relative deviation does not accord with the preset condition, adjusting the reaction rule in the reaction rule set, and recalculating the first relative difference value according to the adjusted reaction rule set until the first relative deviation accords with the preset condition.

Wherein calculating a first relative deviation from a first molecular composition of the device product and a second molecular composition of the actual product preferably comprises: acquiring the types of single molecules in the first molecular composition to form a first set; acquiring the species of single molecules in the second molecular composition to form a second set; determining whether the second set is a subset of the first set; if the second set is not the subset of the first set, acquiring a pre-stored relative deviation value which does not meet a preset condition as the first relative deviation value; if the second set is a subset of the first set, calculating a first relative deviation by:

determining a first relative deviation as a ratio of the number of species of the portion of the molecular composition of the predicted product in which a single molecule is not in the second set to the total number of species of a single molecule in the molecular composition of the predicted product;

for example, the first relative deviation is calculated by the following formula:

wherein x is₁Is the first relative deviation, M is the first set, M₁Is a collection of species compositions of single molecules in the molecular composition of the sample material, M₂Is a collection of species constituents of a single molecule in the molecular composition of the intermediate product, M₃For the second set, card represents the number of elements in the set.

Wherein, using the sample material information to train the reaction rate algorithm preferably comprises: respectively calculating the reaction rate of the reaction path corresponding to each molecule in the molecular composition of the sample raw material according to the reaction rate algorithm; obtaining the predicted content of each molecule in the predicted product corresponding to the sample raw material according to the molecular content of each molecule in the sample raw material and the reaction rate corresponding to the reaction path of the molecule; calculating a second relative deviation according to the predicted content of each molecule in the predicted product and the actual content of each molecule in the actual product; if the second relative deviation meets a preset condition, fixing the reaction rate algorithm; and if the second relative deviation does not accord with the preset condition, adjusting parameters in the reaction rate algorithm, and recalculating the second relative deviation according to the adjusted reaction rate algorithm until the second relative deviation accords with the preset condition.

Wherein, according to the reaction rate algorithm, the reaction rate of the reaction path corresponding to each molecule in the molecular composition of the sample raw material is respectively calculated, and preferably includes: calculating the reaction rate of each reaction path according to the reaction rate constant in the reaction rate algorithm; wherein the reaction rate constant is determined based on a transition state theoretical calculation method;

for example, the reaction rate constant is determined according to the following calculation formula:

wherein k is a reaction rate constant, k_BBoltzmann constant, h is planckian constant, R is ideal gas constant, E is temperature value of environment where reaction path is located, exp is exponential function with natural constant as base, Δ S is entropy change before and after reaction corresponding to reaction rule corresponding to reaction path, Δ E is reaction energy barrier corresponding to reaction rule corresponding to reaction path,

and the catalyst activity factor, P is the pressure value of the environment where the reaction path is located, and alpha is the pressure influence factor corresponding to the reaction rule corresponding to the reaction path.

Among them, the kind of the petroleum processing apparatus preferably includes: catalytic cracker, delayed coking unit, residue hydrogenation unit, hydrocracking unit, diesel oil hydrogenation modification unit, diesel oil hydrogenation refining unit, gasoline hydrogenation refining unit, catalytic reforming unit and alkylation unit; wherein each petroleum processing plant corresponds to a set of reaction rules.

Wherein the reaction conditions preferably include: reaction temperature, reaction pressure and space velocity.

The present invention also provides a product simulation system for a petroleum processing plant, the system comprising:

an acquisition unit for determining the molecular composition of the feedstock;

a first processing unit for inputting the molecular composition of the feedstock into a pre-trained product prediction model; and acquiring the yield of the target product under different reaction conditions output by the product prediction model, wherein the product prediction model corresponds to the type of the petroleum processing device.

Wherein the system further comprises: a second processing unit for querying the yield of the target product under different reaction conditions; determining a reaction condition corresponding to the maximum yield value of the target product; and/or determining the reaction condition corresponding to the minimum yield value of the target product.

Wherein the system further comprises: the model training unit is used for establishing a product prediction model; acquiring sample raw material information of sample raw materials; training a reaction rule set included by the product prediction model by using the sample raw material information, and fixing the trained reaction rule set; training a reaction rate algorithm included in the product prediction model by using the sample raw material information, and fixing the trained reaction rate algorithm to obtain the trained product prediction model, wherein the product prediction model comprises: a reaction rule set and a reaction rate algorithm; the set of reaction rules includes a plurality of reaction rules.

The invention also provides product simulation equipment of the petroleum processing device, which comprises a processor and a memory; the processor is used for executing the product simulation program of the petroleum processing device stored in the memory so as to realize the product simulation method of the petroleum processing device.

The present invention also provides a storage medium storing one or more programs executable by one or more processors to implement the above-described product simulation method for a petroleum processing plant.

The invention has the following beneficial effects:

the invention correspondingly trains a product prediction model for each petroleum processing device in advance, and predicts the yield of the target product of the corresponding petroleum processing device under different reaction conditions through the corresponding product prediction model. The method has the advantages of high accuracy, short time and low cost.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the invention without limiting the invention. In the drawings:

FIG. 1 is a flow diagram of a product simulation method for a petroleum processing plant according to one embodiment of the present invention;

FIG. 2 is a flowchart of the steps for training a artifact prediction model, according to one embodiment of the present invention;

FIG. 3 is a flow chart of the training steps of a reaction rule set according to one embodiment of the present invention;

FIG. 4 is a flow chart of the training steps of the reaction rate algorithm according to one embodiment of the present invention;

FIG. 5 is a flowchart of the steps for property determination according to one embodiment of the present invention;

FIG. 6 is a flowchart of the steps for calculating the physical properties of a single molecule, according to one embodiment of the present invention;

FIG. 7 is a flowchart of the steps for training a physical property calculation model, according to one embodiment of the present invention;

FIG. 8 is a block diagram of a product simulation system of a petroleum processing plant according to one embodiment of the present invention;

fig. 9 is a structural view of a product simulation apparatus of a petroleum processing plant according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail below with reference to the accompanying drawings and specific embodiments.

According to an embodiment of the present invention, a method of product simulation for a petroleum processing plant is provided. Fig. 1 is a flowchart illustrating a product simulation method of a petroleum processing plant according to an embodiment of the present invention.

Step S110, the molecular composition of the feedstock is determined.

The molecular composition of the starting material is information on the various molecules (single molecules) in the starting material. For example: the starting material comprises a single molecule, a kind of single molecule, a volume, a content, etc. of each single molecule.

In this example, the molecular composition of the feedstock can be determined by one or more of comprehensive two-dimensional gas chromatography, quadrupole gas chromatography-mass spectrometer detection, gas chromatography/field ionization-time-of-flight mass spectrometry detection, gas chromatography, near infrared spectroscopy, nuclear magnetic resonance spectroscopy, raman spectroscopy, fourier transform ion cyclotron resonance mass spectrometry, electrostatic field orbitrap mass spectrometry, and ion mobility mass spectrometry.

Of course, the molecular composition of the starting material can also be determined in other ways. For example: the molecular composition of the feedstock was determined by ASTM D2425, SH/T0606, ASTM D8144-18.

In this example, individual single molecules (molecules) in the molecular composition were characterized using a SOL-Oriented Lumping (Structure-Oriented Lumping) based molecular characterization method. Among them, the SOL-based molecular characterization method can characterize the structure of hydrocarbon molecules using 24 groups. A group is a characteristic structure of a portion of a molecule, and each group generally undergoes a chemical reaction as a whole.

Further, the SOL belongs to the lump on the molecular scale, the number of molecules in an actual system is reduced from millions to thousands, and the complexity of physical property detection is greatly reduced. The SOL-based molecular characterization method can represent not only alkanes, cycloalkanes, up to complex aromatic structures containing 50-60 carbon atoms, but also alkenes or cycloalkenes as intermediate products or secondary reaction products, and also heteroatom compounds containing sulfur, nitrogen, oxygen, and the like are considered.

Step S120, inputting the molecular composition of the raw material into a pre-trained product prediction model; wherein the product prediction model corresponds to a type of petroleum processing plant.

The types of petroleum processing equipment include, but are not limited to: catalytic cracker, delayed coking unit, residue hydrogenation unit, hydrocracking unit, diesel oil hydrogenation modification unit, diesel oil hydrogenation refining unit, gasoline hydrogenation refining unit, catalytic reforming unit and alkylation unit; wherein each petroleum processing plant corresponds to a set of reaction rules.

And the product prediction model is used for determining the yield of the product of the raw material under different reaction conditions according to the molecular composition of the raw material. Further, a product prediction model is used for determining the content of various product molecules of the raw material under different reaction conditions according to the molecular composition of the raw material.

The steps for training the product prediction model will be described later, and are not described herein.

And step S130, obtaining the yield of the target product output by the product prediction model under different reaction conditions.

In this embodiment, the reaction conditions include: reaction temperature, reaction pressure and space velocity.

Further, the yields of the target product under different reaction conditions can be represented by a spatial coordinate system, such as: the horizontal axis and the vertical axis are reaction conditions, which may be two of reaction temperature, reaction pressure and space velocity, and the vertical axis may be the yield of the target product.

Specifically, the content of each product molecule output by the product prediction model under different reaction conditions can be obtained; and obtaining the content of the target product molecules under different reaction conditions from the content of each product molecule under different reaction conditions.

In this embodiment, after obtaining the yields of the target products under different reaction conditions output by the product prediction model, the yields of the target products under different reaction conditions may also be queried; determining a reaction condition corresponding to the maximum yield value of the target product; and/or determining the reaction condition corresponding to the minimum yield value of the target product.

In this embodiment, after obtaining the yields of the target products under different reaction conditions output by the product prediction model, the physical properties of the target products under different reaction conditions may also be determined according to the molecular compositions of the target products under different reaction conditions.

The physical properties of the target product under different reaction conditions will be described later, and will not be described herein.

In this embodiment, a product prediction model is trained for each petroleum processing apparatus in advance, and the yield of the target product of the corresponding petroleum processing apparatus under different reaction conditions is predicted by the corresponding product prediction model. The method has the advantages of high accuracy, short time and low cost.

The steps for training the artifact prediction model are further described below. FIG. 2 is a flowchart illustrating steps for training a product prediction model according to an embodiment of the present invention.

And step S210, establishing a product prediction model.

The product prediction model comprises: a set of reaction rules and a reaction rate algorithm. Wherein the set of reaction rules includes a plurality of reaction rules.

And correspondingly establishing a product prediction model according to the type of the petroleum processing device. The reaction rule sets corresponding to different types of petroleum processing equipment may be different. The reaction rate algorithms for different types of petroleum processing equipment may also be different.

Further, the product prediction model corresponding to the kind of petroleum processing apparatus includes: a set of reaction rules and a reaction rate algorithm corresponding to the category of petroleum processing equipment. Wherein the set of reaction rules includes: a plurality of reaction rules corresponding to the type of petroleum processing equipment.

Step S220, sample material information of the sample material is obtained.

Sample material information of the sample material includes: the molecular composition of the sample raw material, the molecular content of each molecule in the sample raw material, the molecular composition of an actual product corresponding to the sample raw material and the actual content of each molecule in the actual product. The actual product is the product obtained after the sample raw material is processed by the petroleum processing device of the kind.

And step S230, training the reaction rule set by using the sample raw material information, and fixing the trained reaction rule set.

One way to train the reaction rule set is given below. It should be understood by those skilled in the art that the manner is only for illustrating the embodiment and is not limited to the embodiment.

FIG. 3 is a flowchart illustrating the training steps of the reaction rule set according to an embodiment of the present invention.

Step S310, processing the molecular composition of the sample raw material according to a reaction rule set to obtain a reaction path corresponding to each molecule in the molecular composition of the sample raw material.

The reaction path is used to indicate a chemical reaction path of an intermediate product obtained by reacting each molecule or a final product obtained by reacting.

And when the reaction path is calculated for the first time, processing the molecular composition of the sample raw material according to a preset reaction rule set to obtain the reaction path corresponding to each molecule in the molecular composition of the sample raw material.

And (3) reacting each molecule in the sample raw material according to the reaction rule in the reaction rule set to obtain a reaction path corresponding to each molecule. After each molecule is subjected to the first reaction to generate an intermediate product, the molecular structure of the intermediate product may meet another reaction rule, and the intermediate product continues to perform subsequent reactions according to the met reaction rule until the molecular structure of the product generated by the molecule does not meet any reaction rule in the reaction rule set, so as to obtain a final product of the molecular reaction, wherein the summary of the reactions is the reaction path of the molecule.

Step S320, obtaining a first molecular composition of the device product according to the reaction path corresponding to each molecule in the molecular composition of the sample raw material.

In the device product, comprising: the sample feedstock, intermediate product, and predicted product.

Step S330, calculating a first relative deviation according to the first molecular composition of the device product and the second molecular composition of the actual product.

Specifically, acquiring the types of single molecules in the first molecular composition to form a first set; acquiring the species of single molecules in the second molecular composition to form a second set; determining whether the second set is a subset of the first set; if the second set is not the subset of the first set, acquiring a pre-stored relative deviation value which does not meet a preset condition as the first relative deviation value; if the second set is a subset of the first set, calculating a first relative deviation by:

wherein x is₁Is the first relative deviation, M is the first set, M₁Is a collection of species compositions of single molecules in the molecular composition of the sample material, M₂Is a collection of species constituents of a single molecule in the molecular composition of the intermediate product, M₃For the second set, card represents the number of elements in the set. M, M₁、M₂The number of elements is the same as that of the elements in N.

The preset conditions comprise: relative deviation range. The two endpoints of the relative deviation range are empirical values or experimentally obtained values. The pre-stored relative deviation value is not within the relative deviation range.

Step S340, judging whether the first relative deviation meets a preset condition; if yes, go to step S350; if not, step S360 is performed.

And if the first relative deviation is within the relative deviation range, judging that the first relative deviation meets the preset condition, otherwise, judging that the first relative deviation does not meet the preset condition.

Step S350, if the first relative deviation meets a preset condition, fixing the reaction rule set.

Step S360, if the first relative deviation does not meet the preset condition, adjusting the reaction rule in the reaction rule set, and jumping to step S310, and recalculating the first relative difference value according to the adjusted reaction rule set until the first relative deviation meets the preset condition.

And S240, training the reaction rate algorithm by using the sample raw material information, and fixing the trained reaction rate algorithm to obtain the trained product prediction model.

One way to train the reaction rate algorithm is given below. It should be understood by those skilled in the art that the manner is only for illustrating the embodiment and is not limited to the embodiment.

FIG. 4 is a flowchart illustrating the training steps of the reaction rate algorithm according to an embodiment of the present invention.

Step S410, respectively calculating the reaction rate corresponding to the reaction path corresponding to each molecule in the molecular composition of the sample raw material according to a reaction rate algorithm.

Specifically, the reaction rate of each reaction path is calculated according to a reaction rate constant in the reaction rate algorithm;

determining the reaction rate constant according to the following calculation formula:

the activity factor of the catalyst is determined by the following formula,p is the pressure value of the environment where the reaction path is located, and alpha is the pressure influence factor corresponding to the reaction rule corresponding to the reaction path.

Specifically, the reaction rate of the reaction path is obtained according to the reaction rate constant and the reaction concentration corresponding to the reaction path. For example: where a reaction rate constant has been determined, the greater the space velocity, the shorter the contact time of the feedstock with the catalyst, the shorter the reaction time of the feedstock, the higher the concentration of reactants in the feedstock, and the higher the reaction rate of the reaction path; conversely, the lower the space velocity, the longer the contact time between the feedstock and the catalyst, the longer the reaction time of the feedstock, the lower the concentration of reactants in the feedstock, and the lower the reaction rate of the reaction path.

In this embodiment, k in the above formula is referred to as a reaction rate constant, and its physical meaning is that its value is equivalent to the reaction rate when all the substances participating in the reaction are at a unit concentration (1mol/L), so it is also referred to as a specific rate of the reaction, and different reactions have different rate constants depending on the reaction temperature, the reaction medium (solvent), the catalyst, etc., and even depending on the shape and properties of the reactor. Regardless of concentration, but is influenced by factors such as temperature, catalyst, surface properties of the solid, and the like. In this embodiment, after calculating the reaction rate constant, the real-time reaction rate of the molecule can be obtained according to the concentration of the molecule, for example, if the reaction rate of the molecule per unit concentration is k, and after confirming the concentration V, the reaction rate of the molecule corresponding to the reaction path is Vk, so as to calculate the reaction rate.

And step S420, obtaining the predicted content of each molecule in the predicted product corresponding to the sample raw material according to the molecular content of each molecule in the sample raw material and the reaction rate corresponding to the reaction path of the molecule.

In this embodiment, the reaction rate corresponding to each reaction path is calculated by a reaction rate calculation method in the product prediction model, and the predicted content of each monomolecular in the predicted product can be calculated by combining the monomolecular content of each monomolecular in the raw material.

For example, assuming that the monomolecular a in the raw material corresponds to 3 reaction paths, the reaction rates corresponding to the 3 reaction paths are known, and as the reaction proceeds, the concentration of the monomolecular a decreases, and the reaction rates corresponding to the 3 reaction paths decrease according to the decreasing rate of the concentration, so that the monomolecular a generates a product according to the rate of the reaction rates of the 3 reaction paths, and according to the above steps, the product obtained by reacting each molecule can be obtained, and the predicted product can be obtained, and when the monomolecular content of each monomolecular in the catalytic reforming raw material is known, the content of each monomolecular in the predicted product can be obtained.

Step S430, calculating a second relative deviation according to the predicted content of each molecule in the predicted product and the actual content of each molecule in the actual product.

The second relative deviation is calculated, for example, as:

second relative deviation ═ (actual content-predicted content)/actual content.

Step S440, judging whether the second relative deviation meets a preset condition; if yes, go to step S450; if not, step S460 is performed.

And if the second relative deviation is within the relative deviation range, judging that the second relative deviation meets the preset condition, otherwise, judging that the second relative deviation does not meet the preset condition.

Step S450, if the second relative deviation meets a preset condition, fixing the reaction rate algorithm.

Step S460, if the second relative deviation does not meet the preset condition, adjusting parameters in the reaction rate algorithm, skipping to step S410, and recalculating the second relative deviation according to the adjusted reaction rate algorithm until the second relative deviation meets the preset condition.

In this embodiment, parameters in the reaction rate algorithm in the reaction rate calculation method corresponding to each reaction path in the product prediction model are adjusted. The accuracy of the reaction rate algorithm in the product prediction model is ensured by feedback adjustment.

The following description will be given to determine the physical properties of the target product under different reaction conditions.

Fig. 5 is a flowchart illustrating the steps of determining physical properties according to an embodiment of the present invention.

Step S510, a target product under each reaction condition is obtained, and various single molecules contained in the target product are determined.

Determining the various single molecules contained in the target product, i.e. determining the molecular composition of the target product.

Step S520, calculating the physical properties of each single molecule in the target product according to the number of groups of each group constituting the single molecule and the contribution value of each group to the physical properties.

Physical properties of single molecules, including but not limited to: density, boiling point, density, octane number. For example: the physical properties of the single molecule may further include: viscosity, solubility parameter, cetane number, unsaturation, and the like.

A group is a part of a molecule, and the group generally participates in a chemical reaction as a whole.

Determining the group contained in each single molecule based on the SOL molecular characterization method; the number of groups per group in each single molecule and the contribution value of each group in the single molecule to the physical properties are determined, respectively. Since the number of physical properties of a single molecule is large, it is necessary to determine the contribution value of each group in the single molecule to each physical property.

The specific way of calculating the physical properties of the single molecule will be described later, and will not be described herein.

Step S530, calculating the physical properties of the target product according to the physical properties of various single molecules in the target product and the content of various single molecules in the target product.

Physical properties of the target product include, but are not limited to: density, cloud point, pour point, aniline point, and octane number. Of course, the physical properties of the target product may include: cetane number, freezing point, cold filter plugging point, etc.

Five ways of calculating the physical properties of the target product are provided below, but those skilled in the art should understand that the following ways are only used for illustrating the present embodiment and are not used for limiting the present embodiment.

In the first embodiment, when the physical property of the target product is density, the density of the target product is calculated by the following calculation formula:

density＝∑(D_i×x_i-volume)；

wherein density is the density of the target product, D_iDensity of the i-th single molecule, x_i-volumeThe content of the i-th single molecule.

In the second aspect, the calculating the physical property of the target product when the physical property of the target product is a cloud point includes:

calculating a cloud point contribution for each single molecule based on the density and boiling point of the single molecule;

calculating the cloud point of the target product according to the cloud point contribution value of each single molecule in the target product and the content of each single molecule in the target product.

In a third aspect, when the physical property of the target product is a pour point, calculating the physical property of the target product includes:

calculating a pour point contribution value for each single molecule based on the density and molecular weight of the single molecule;

calculating the pour point of the target product based on the pour point contribution value of each single molecule in the target product and the content of each single molecule in the target product.

In a fourth embodiment, when the physical property of the target product is an aniline point, calculating the physical property of the target product includes:

calculating the aniline point contribution value of each single molecule according to the density and boiling point of the single molecule;

and calculating the aniline point of the target product according to the aniline point contribution value of each single molecule in the target product and the content of each single molecule in the target product.

And in the fifth mode, when the physical property of the target product is the octane number, calculating the octane number of the target product by the following calculation formula:

wherein ON is the octane number of the target product, HISQFG is a molecular set, H is a molecular set of normal paraffin, I is a molecular set of isoparaffin, S is a molecular set of cycloalkane, Q is a molecular set of olefin, F is a molecular set of aromatic hydrocarbon, G is a molecular set of oxygen-containing compound, and upsilon is_iThe content of each molecule in the target product; upsilon is_H、υ_I、υ_S、υ_Q、υ_F、υ_GRespectively representing the total content of normal paraffin, the total content of isoparaffin, the total content of cyclane, the total content of olefin, the total content of aromatic hydrocarbon and the total content of oxygen-containing compounds in the target product; beta is a_iRegression parameters for each molecule in the target product; ON_iAn octane number for each molecule in the target product; c_HRepresenting the interaction coefficient of the normal alkane with other molecules; c_IRepresenting the interaction coefficient of the isoparaffin with other molecules; c_SRepresenting the coefficient of interaction of cycloalkanes with other molecules; c_QRepresenting the coefficient of interaction of the olefin with other molecules; c_FRepresenting the interaction coefficient of the aromatic hydrocarbon with other molecules; c_GRepresenting the interaction coefficient of the oxygen-containing compound and other molecules;

a first constant coefficient between the normal paraffin and the isoparaffin,

A first constant coefficient between n-alkane and cycloalkane,

A first constant coefficient between the normal paraffin and the olefin,

A first constant coefficient between n-alkane and aromatic hydrocarbon,

A first constant coefficient between the normal alkane and the oxygen-containing compound,

A first constant coefficient between isoparaffin and cycloalkane,

A first constant coefficient between the isoparaffin and the olefin,

A first constant coefficient between isoparaffin and aromatic hydrocarbon,

A first constant coefficient between the isoparaffin and the oxygen-containing compound,

Represents cycloalkane withA first constant coefficient between olefins,

A first constant coefficient between a cycloalkane and an aromatic hydrocarbon,

A first constant coefficient representing the ratio between the cycloalkane and the oxygen-containing compound,

A first constant coefficient between olefin and aromatic hydrocarbon,

A first constant coefficient between the olefin and the oxygen-containing compound,

A first constant coefficient between the aromatic hydrocarbon and the oxygen-containing compound,

A bi-constant coefficient between the normal paraffin and the isoparaffin,

A second constant coefficient between n-alkane and cycloalkane,

A second constant coefficient between the normal paraffin and the olefin,

A second constant coefficient between the normal paraffin and the aromatic hydrocarbon,

A second constant coefficient between the normal alkane and the oxygen-containing compound,

A second constant coefficient between isoparaffin and cycloalkane,

A second constant coefficient between the isoparaffin and the olefin,

A second constant coefficient between isoparaffin and aromatic hydrocarbon,

A second constant coefficient between the isoparaffin and the oxygen-containing compound,

A second constant coefficient between cycloalkane and olefin,

A second constant coefficient between the cycloalkane and the aromatic hydrocarbon,

A second constant coefficient representing the ratio between the cycloalkane and the oxygen-containing compound,

A second constant coefficient between olefin and aromatic hydrocarbon,

A second constant coefficient between the olefin and the oxygen-containing compound,

Represents a second constant coefficient between the aromatic hydrocarbon and the oxygen-containing compound; wherein the octane number comprises: research octane number and motor octane number.

In this embodiment, before calculating the physical properties of the single molecule, the number of groups of each group constituting the single molecule may be compared with the molecular information of the template single molecule with known physical properties pre-stored in the database; wherein the molecular information comprises: the number of groups of each group constituting a single molecule of the template; judging whether a template single molecule identical to the single molecule exists; if a template monomolecular identical to the monomolecular exists, outputting the physical property of the template monomolecular as the physical property of the monomolecular; and if the template single molecule identical to the single molecule does not exist, calculating the physical property of the single molecule. Further, if the kind of the group constituting the template single molecule is the same as that of the group constituting the single molecule and the number of groups per group of the template single molecule is the same as that of the group corresponding to the single molecule, it is determined that the template single molecule is the same as the single molecule, and otherwise, it is determined that the template single molecule is different from the single molecule.

In this embodiment, after the number of groups of each group constituting a single molecule is obtained, it is determined whether the structure of the single molecule is stored in the database by comparing the corresponding number of groups, and after the occurrence of a template single molecule identical to the single molecule is determined, the physical properties of the single molecule are directly output, so that the calculation efficiency of the physical properties of the single molecule is improved, and the calculation amount is reduced.

The calculation of the physical properties of the single molecules is further described below.

FIG. 6 is a flowchart illustrating steps for calculating the physical properties of a single molecule according to an embodiment of the present invention.

Step S610, for each single molecule in the target product, acquiring the number of groups of each group constituting the single molecule, and acquiring a contribution value of each group to the physical property.

Further, the kind of the group contained in a single molecule is determined, the number of the group of each kind of the group is determined, and the contribution value of each kind of the group to each physical property of the target product is obtained.

Step S620 is to input the number of groups of each group constituting the single molecule and the contribution value of each group to the physical property into a physical property calculation model trained in advance, and to obtain the physical property of the single molecule output by the physical property calculation model.

In this example, a physical property calculation model for calculating a physical property of a single molecule based on the number of groups of each group contained in the single molecule and a value of contribution of each group to the physical property.

Further, the number of groups of each kind of group of a single molecule is obtained, and the contribution value of each kind of group to each physical property of the target product is obtained.

The steps for training the physical property calculation model are further described below.

FIG. 7 is a flowchart illustrating steps for training a physical property computation model according to an embodiment of the present invention.

Step S710, a monomolecular physical property calculation model is constructed.

The physical property calculation model includes: contribution of each group to physical properties. The contribution value is an adjustable value, and the contribution value is an initial value when training for the first time. Further, the physical property calculation model includes: contribution of each group to each physical property.

Two types of physical property calculation models that can be used for different physical properties are given below. It should be understood by those skilled in the art that the following two physical property calculation models are only illustrative of the present embodiment and are not intended to limit the present embodiment.

Model one: a physical property calculation model shown below was established:

wherein f is the physical property of the single molecule, and n is_iNumber of groups of i-th group,. DELTA.f_iThe value of contribution of the i-th group to the physical property, and a is a correlation constant.

Groups constituting a single molecule can be further classified into multi-stage groups. Further, defining a primary group and a multi-order group among all groups of the single molecule; wherein all groups constituting a single molecule are taken as primary groups; a plurality of groups which exist simultaneously and contribute to the common existence of the same physical property are used as a multi-stage group, and the number of the plurality of groups is used as the level of the multi-stage group. Further, when the groups are present in different molecules independently, the physical properties are influenced to a certain degree, and when the groups are present in one molecule at the same time, the contribution value to the physical properties is fluctuated to a certain degree in addition to the original contribution value to the physical properties. The manner of dividing the above-mentioned multilevel groups may also be divided by a bonding force interval to which the chemical bonding force between the groups belongs. The chemical bond strength may have different influences on different physical properties, and specifically, the influence on the physical properties may be classified according to the molecular stability.

For example: for boiling point, in the SOL-based molecular characterization method, 24 groups are all used as primary groups; of the 24 groups, the presence of one or more of N6, N5, N4, N3, me, AA, NN, RN, NO, RO, KO simultaneously contributes to boiling point. When dividing a single molecule of groups, all groups forming the single molecule are used as first-order groups, whether a plurality of groups which can contribute to the boiling point together exist in all the groups of the single molecule or not is checked, if the groups exist, a plurality of groups which can contribute to the boiling point together exist together and are used as multi-order groups, such as: if both N6 and N4 are present in the single molecule, then the number of groups that contribute to boiling point co-presence when present is two, then the combination of N6 and N4 is considered to be a secondary group.

Model two: based on the divided multilevel groups, a physical property calculation model as described below can be established:

wherein f is the physical property of the single molecule, and m is_1iIs the number of groups of the i-th group in the primary group,. DELTA.f_1iM is the value of the contribution of the i-th group in the primary group to the physical properties_2jIs the number of groups of the jth group in the secondary group,. DELTA.f_2jIs the contribution value of the jth group in the secondary group to the physical property;m_Nlis the number of groups of the group I in the N-th group,. DELTA.f_NlIs the contribution value of the first group in the N-grade groups to physical properties; a is a correlation constant; n is a positive integer greater than or equal to 2.

Step S720, acquiring the number of groups of each group forming a single molecule of the sample; wherein the physical properties of the sample single molecule are known.

A training sample set is preset. A plurality of sample single molecule information is included in the training sample set. Sample single molecule information including, but not limited to: the number of groups of each group constituting a single molecule of the sample, and the physical properties of the single molecule of the sample.

In step S730, the number of groups of each group constituting a single molecule of the sample is input to the physical property calculation model.

Step S740, obtaining the predicted physical properties of the sample single molecule output by the physical property calculation model.

Step S750, judging whether the deviation value between the predicted physical property and the known physical property is smaller than a preset deviation threshold value; if yes, go to step S760; if not, step S770 is executed.

If the physical property of the sample single molecule is multiple, the predicted physical property of the sample single molecule output by the physical property calculation model is multiple, at this time, a deviation value between each predicted physical property and the corresponding known physical property is calculated, whether the deviation value between each predicted physical property and the corresponding known physical property is smaller than a preset deviation value is judged, and if yes, the step S760 is executed; if not, step S770 is executed.

In this embodiment, a relative deviation value or an absolute deviation value of the predicted physical property and the corresponding known physical property may be calculated.

Step S760 of determining that the property calculation model converges if a deviation value between the predicted property and the known property is smaller than a preset deviation threshold value, acquiring a contribution value of each group to the property in the converged property calculation model, and storing the contribution value of the group to the property.

Since there may be a plurality of types of physical properties of a single molecule, the contribution value of each group to each physical property can be obtained in a converged physical property calculation model.

In this embodiment, the physical property calculation model is constructed, and the physical property calculation model is trained, so that the physical property calculation model converges, that is, the contribution value of each group to the physical property in the physical property calculation model is trained, and the contribution value of each group to the physical property is obtained.

For each group, storing the contribution value of the group to each physical property, so that when the physical property of a single molecule is calculated later, the contribution value of each group in the single molecule to the physical property to be known can be obtained, and the number of groups of each group in the single molecule and the contribution value of each group to the physical property to be known are used as the input of a physical property calculation model, the physical property calculation model uses the number of groups of each group in the single molecule as a model variable, and uses the contribution value of each group to the physical property to be known as a model parameter (replacing the adjustable contribution value of each group in the physical property calculation model to the physical property), and the physical property to be known is calculated.

Step S770, if the deviation value between the predicted physical property and the known physical property is equal to or greater than the deviation threshold value, adjusting the contribution value of each group in the physical property calculation model to the physical property, and proceeding to step S750 until the physical property calculation model converges.

In addition to the general-purpose physical property calculation model described above, a physical property calculation model may be constructed for each physical property depending on the type of physical property.

For example: the boiling point of the single molecule was calculated according to the following physical property calculation model:

wherein T is the boiling point of the single molecule, SOL is the monomolecular vector converted according to the number of GROUPs of each GROUP constituting the single molecule, GROUP₁₁GROUP, a first contribution vector converted from the contribution of the primary GROUP to the boiling point₁₂GROUP, a second contribution vector converted from the contribution of the secondary GROUP to the boiling point_1NThe N contribution value vector is obtained by converting the contribution value of the N-level group to the boiling point, Numh is the number of atoms except hydrogen atoms in a single molecule, d is a first preset constant, b is a second preset constant, and c is a third preset constant; n is a positive integer greater than or equal to 2.

b. c and d may be empirical values or values obtained by experiment.

A monomolecular vector converted according to the number of groups of each group constituting the monomolecular, comprising: taking the number of species of groups constituting the single molecule as the dimension of the single molecule vector; the number of groups per said group is taken as the element value of the corresponding dimension in said single molecular vector.

The first contribution value vector obtained by converting the contribution value of each primary group of the single molecule to the boiling point comprises: taking the number of species of the primary group as the dimension of the first contribution vector; and taking the contribution value of each primary group as the element value of the corresponding dimension in the first contribution value vector. And converting a second contribution value vector according to the contribution values of each secondary group of the single molecule to the boiling point respectively, wherein the second contribution value vector comprises: taking the number of species of secondary groups as the dimension of the second contribution vector; and taking the contribution value of each secondary group as the element value of the corresponding dimension in the second contribution value vector. By analogy, the Nth contribution value vector obtained by converting the contribution values of each N-grade group of the single molecule to the boiling point respectively comprises the following components: taking the number of species of the N-th order group as the dimension of the Nth contribution value vector; and taking the contribution value of each N-grade group as the element value of the corresponding dimension in the Nth contribution value vector.

For another example: the density of the single molecules was calculated according to the following physical property calculation model:

wherein D is the density of the single molecule, SOL is a single molecular vector converted according to the number of GROUPs of each GROUP constituting the single molecule, GROUP₂₁GROUP is the vector of N +1 contribution converted from the contribution of the primary GROUP to the density₂₂GROUP is the vector of N +2 contribution converted from the contribution of secondary GROUPs to the density_2NThe vector of the 2N contribution value is obtained by converting the contribution value of the N-grade group to the density, and e is a fourth preset constant; n is a positive integer greater than or equal to 2.

e may be an empirical value or a value obtained by experiment.

The N +1 th contribution value vector obtained by converting the contribution values of the primary groups of the single molecule to the density respectively comprises: taking the number of species of the primary group as the dimension of the N +1 th contribution vector; and taking the contribution value of each primary group as the element value of the corresponding dimension in the N +1 th contribution value vector. And (3) converting the N +2 contribution value vector obtained according to the contribution value of each secondary group of the single molecule to the density respectively, wherein the vector comprises: taking the number of species of secondary groups as the dimension of the N +2 contribution vector; and taking the contribution value of each secondary group as the element value of the corresponding dimension in the N +2 th contribution value vector. By analogy, the 2N contribution value vector obtained by converting the contribution values of each N-level group of the single molecule to the density respectively comprises: taking the number of species of the N-th order group as the dimensionality of the 2N contribution vector; and taking the contribution value of each N-grade group as the element value of the corresponding dimension in the 2N contribution value vector.

The following steps are repeated: the octane number of a single molecule was calculated according to the following physical property calculation model:

X＝SOL×GROUP₃₁+SOL×GROUP₃₂+......+SOL×GROUP_3N+h；

wherein X is the octane number of the single molecule, SOL is a single molecular vector converted according to the number of GROUPs of each GROUP constituting the single molecule, GROUP₃₁GROUP is a 2N +1 contribution vector converted from the contribution of the primary GROUP to the octane number₃₂GROUP is a 2N +2 contribution vector converted from the contribution of the secondary GROUP to the octane number_3NThe 3N contribution value vector is obtained by converting the contribution value of the N-grade group to the octane number; n is a positive integer greater than or equal to 2; h is a fifth predetermined constant. h is an empirical value or a value obtained by experiment.

The 2N +1 th contribution value vector obtained by converting the contribution value of each primary group of the single molecule to the octane number respectively comprises the following components: taking the number of species of the primary group as the dimensionality of the 2N +1 th contribution vector; and taking the contribution value of each primary group as the element value of the corresponding dimension in the 2N +1 th contribution value vector. The 2N +2 contribution value vector obtained by converting the contribution value of each secondary group of the single molecule to the octane number respectively comprises the following components: taking the number of species of secondary groups as the dimensionality of the 2N +2 contribution vector; and taking the contribution value of each secondary group as the element value of the corresponding dimension in the 2N +2 contribution value vector. By analogy, the 3N contribution value vector obtained by converting the contribution values of each N-grade group of the single molecule to the octane number respectively comprises the following steps: taking the number of species of the N-th order group as the dimension of the 3 Nth contribution value vector; and taking the contribution value of each N-grade group as the element value of the corresponding dimension in the 3 Nth contribution value vector.

The invention also provides a product simulation system of the petroleum processing device. Fig. 8 is a block diagram of a product simulation apparatus of a petroleum processing system according to an embodiment of the present invention.

Referring to fig. 8, the simulation system includes: an acquisition unit 11 and a first processing unit 12.

In the present embodiment, an obtaining unit 11 for determining the molecular composition of the raw material;

in this embodiment, the first processing unit 12 is configured to input the molecular composition of the raw material into a pre-trained product prediction model; and acquiring the yield of the target product under different reaction conditions output by the product prediction model, wherein the product prediction model corresponds to the type of the petroleum processing device.

Wherein, the system still includes: and the third processing unit is used for determining the physical properties of the target product under different reaction conditions according to the molecular composition of the target product under different reaction conditions.

The first processing unit 12 is specifically configured to obtain the content of each product molecule output by the product prediction model under different reaction conditions; and obtaining the content of the target product molecules under different reaction conditions from the content of each product molecule under different reaction conditions.

The acquiring unit 11 is specifically configured to determine the molecular composition of the raw material by one or more of a comprehensive two-dimensional gas chromatography, a quadrupole gas chromatography-mass spectrometer detection method, a gas chromatography/field ionization-time-of-flight mass spectrometry detection method, a gas chromatography, a near-infrared spectroscopy, a nuclear magnetic resonance spectroscopy, a raman spectroscopy, a fourier transform ion cyclotron resonance mass spectrometry, an electrostatic field orbitrap mass spectrometry, and an ion mobility mass spectrometry.

Wherein, the sample material information of the sample material comprises: the molecular composition of the sample raw material, the molecular content of each molecule in the sample raw material, the second molecular composition of the actual product corresponding to the sample raw material and the actual content of each molecule in the actual product.

The model training unit is specifically used for processing the molecular composition of the sample raw material according to a preset reaction rule set to obtain a reaction path corresponding to each molecule in the molecular composition of the sample raw material; obtaining a first molecular composition of a device product according to a reaction path corresponding to each molecule in the molecular composition of the sample raw material; in the device product, comprising: the sample feedstock, intermediate product, and predicted product; calculating a first relative deviation from a first molecular composition of the device product and a second molecular composition of the actual product; if the first relative deviation meets a preset condition, fixing the reaction rule set; and if the first relative deviation does not accord with the preset condition, adjusting the reaction rule in the reaction rule set, and recalculating the first relative difference value according to the adjusted reaction rule set until the first relative deviation accords with the preset condition.

The model training unit is specifically used for acquiring the types of the single molecules in the first molecular composition to form a first set; acquiring the species of single molecules in the second molecular composition to form a second set; determining whether the second set is a subset of the first set; if the second set is not the subset of the first set, acquiring a pre-stored relative deviation value which does not meet a preset condition as the first relative deviation value; if the second set is a subset of the first set, calculating a first relative deviation by:

The model training unit is specifically used for respectively calculating the reaction rate of the reaction path corresponding to each molecule in the molecular composition of the sample raw material according to the reaction rate algorithm; obtaining the predicted content of each molecule in the predicted product corresponding to the sample raw material according to the molecular content of each molecule in the sample raw material and the reaction rate corresponding to the reaction path of the molecule; calculating a second relative deviation according to the predicted content of each molecule in the predicted product and the actual content of each molecule in the actual product; if the second relative deviation meets a preset condition, fixing the reaction rate algorithm; and if the second relative deviation does not accord with the preset condition, adjusting parameters in the reaction rate algorithm, and recalculating the second relative deviation according to the adjusted reaction rate algorithm until the second relative deviation accords with the preset condition.

The model training unit is specifically used for calculating the reaction rate of each reaction path according to the reaction rate constant in the reaction rate algorithm; wherein the reaction rate constant is determined according to the following calculation formula:

wherein k is a reaction rate constant, k_BIs the boltzmann constant, and is,h is a Planck constant, R is an ideal gas constant, E is a temperature value of the environment where the reaction path is located, exp is an exponential function with a natural constant as a base, Delta S is an entropy change before and after the reaction corresponding to the reaction rule corresponding to the reaction path, Delta E is a reaction energy barrier corresponding to the reaction rule corresponding to the reaction path,

Wherein the petroleum processing device comprises the following types: catalytic cracker, delayed coking unit, residue hydrogenation unit, hydrocracking unit, diesel oil hydrogenation modification unit, diesel oil hydrogenation refining unit, gasoline hydrogenation refining unit, catalytic reforming unit and alkylation unit; wherein each petroleum processing plant corresponds to a set of reaction rules.

Wherein the reaction conditions comprise: reaction temperature, reaction pressure and space velocity.

The present embodiment provides a product simulation apparatus for a petroleum processing plant. Fig. 9 is a block diagram of a product simulation apparatus of a petroleum processing plant according to an embodiment of the present invention.

In this embodiment, the product simulation device of the petroleum processing plant includes, but is not limited to: a processor 810, a memory 820.

The processor 810 is configured to execute a product simulation program of the petroleum processing device stored in the memory 820 to implement the product simulation method of the petroleum processing device.

Specifically, the processor 810 is configured to execute a product simulation program of a petroleum processing plant stored in the memory 820 to implement the steps of: determining the molecular composition of the feedstock; inputting the molecular composition of the raw material into a pre-trained product prediction model; wherein the product prediction model corresponds to a type of petroleum processing plant; and obtaining the yield of the target product under different reaction conditions output by the product prediction model.

After obtaining the yields of the target products under different reaction conditions, which are output by the product prediction model, the method further comprises the following steps: inquiring the yield of the target product under different reaction conditions; determining a reaction condition corresponding to the maximum yield value of the target product; and/or determining the reaction condition corresponding to the minimum yield value of the target product.

After obtaining the yields of the target products under different reaction conditions, which are output by the product prediction model, the method further comprises the following steps: and determining the physical properties of the target product under different reaction conditions according to the molecular composition of the target product under different reaction conditions.

Obtaining the yield of the target product output by the product prediction model under different reaction conditions, wherein the yield of the target product output by the product prediction model under different reaction conditions comprises the following steps: obtaining the content of each product molecule output by the product prediction model under different reaction conditions; and obtaining the content of the target product molecules under different reaction conditions from the content of each product molecule under different reaction conditions.

Wherein determining the molecular composition of the feedstock comprises: determining the molecular composition of the feedstock by one or more of comprehensive two-dimensional gas chromatography, quadrupole gas chromatography-mass spectrometry detection, gas chromatography/field ionization-time-of-flight mass spectrometry detection, gas chromatography, near infrared spectroscopy, nuclear magnetic resonance spectroscopy, raman spectroscopy, fourier transform ion cyclotron resonance mass spectrometry, electrostatic field orbitrap mass spectrometry, and ion mobility mass spectrometry.

Wherein the step of training the product prediction model comprises: establishing a product prediction model; wherein the product prediction model comprises: a reaction rule set and a reaction rate algorithm; the set of reaction rules comprises a plurality of reaction rules; acquiring sample raw material information of sample raw materials; training the reaction rule set by using the sample raw material information, and fixing the trained reaction rule set; and training the reaction rate algorithm by using the sample raw material information, and fixing the trained reaction rate algorithm to obtain the trained product prediction model.

Wherein, the sample material information of the sample material comprises: the molecular composition of the sample raw material, the molecular content of each molecule in the sample raw material, the molecular composition of an actual product corresponding to the sample raw material and the actual content of each molecule in the actual product.

Wherein training the set of reaction rules using the sample material information comprises: processing the molecular composition of the sample raw material according to a preset reaction rule set to obtain a reaction path corresponding to each molecule in the molecular composition of the sample raw material; obtaining a first molecular composition of a device product according to a reaction path corresponding to each molecule in the molecular composition of the sample raw material; in the device product, comprising: the sample feedstock, intermediate product, and predicted product; calculating a first relative deviation from a first molecular composition of the device product and a second molecular composition of the actual product; if the first relative deviation meets a preset condition, fixing the reaction rule set; and if the first relative deviation does not accord with the preset condition, adjusting the reaction rule in the reaction rule set, and recalculating the first relative difference value according to the adjusted reaction rule set until the first relative deviation accords with the preset condition.

Wherein calculating a first relative deviation from a first molecular composition of the device product and a second molecular composition of the actual product comprises: acquiring the types of single molecules in the first molecular composition to form a first set; acquiring the species of single molecules in the second molecular composition to form a second set; determining whether the second set is a subset of the first set; if the second set is not the subset of the first set, acquiring a pre-stored relative deviation value which does not meet a preset condition as the first relative deviation value; if the second set is a subset of the first set, calculating a first relative deviation by:

wherein x is₁In order to be able to determine the first relative deviation,m is the first set, M₁Is a collection of species compositions of single molecules in the molecular composition of the sample material, M₂Is a collection of species constituents of a single molecule in the molecular composition of the intermediate product, M₃For the second set, card represents the number of elements in the set.

Wherein training the reaction rate algorithm using the sample material information comprises: respectively calculating the reaction rate of the reaction path corresponding to each molecule in the molecular composition of the sample raw material according to the reaction rate algorithm; obtaining the predicted content of each molecule in the predicted product corresponding to the sample raw material according to the molecular content of each molecule in the sample raw material and the reaction rate corresponding to the reaction path of the molecule; calculating a second relative deviation according to the predicted content of each molecule in the predicted product and the actual content of each molecule in the actual product; if the second relative deviation meets a preset condition, fixing the reaction rate algorithm; and if the second relative deviation does not accord with the preset condition, adjusting parameters in the reaction rate algorithm, and recalculating the second relative deviation according to the adjusted reaction rate algorithm until the second relative deviation accords with the preset condition.

Wherein, according to the reaction rate algorithm, respectively calculating the reaction rate of the reaction path corresponding to each molecule in the molecular composition of the sample raw material, and comprises: calculating the reaction rate of each reaction path according to the reaction rate constant in the reaction rate algorithm; wherein,

wherein k is a reaction rate constant, k_BBoltzmann constant, h is Planck constant, R is ideal gas constant, E is temperature value of environment where reaction path is located, exp is exponential function with natural constant as base, Δ S is entropy change before and after reaction corresponding to reaction rule corresponding to reaction path, and Δ E is entropy change before and after reaction corresponding to reaction rule corresponding to reaction pathA reaction energy barrier corresponding to the reaction rule corresponding to the path,

The embodiment of the invention also provides a storage medium. The storage medium herein stores one or more programs. Among others, the storage medium may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as read-only memory, flash memory, a hard disk, or a solid state disk; the memory may also comprise a combination of memories of the kind described above.

When the one or more programs in the storage medium are executable by the one or more processors to implement the above-described method for product simulation for a petroleum processing plant.

Specifically, the processor is configured to execute a product simulation program of a petroleum processing plant stored in the memory to perform the steps of: determining the molecular composition of the feedstock; inputting the molecular composition of the raw material into a pre-trained product prediction model; wherein the product prediction model corresponds to a type of petroleum processing plant; and obtaining the yield of the target product under different reaction conditions output by the product prediction model.

wherein x is₁Is the first relative deviation, M is the first set, M₁Is a collection of species compositions of single molecules in the molecular composition of the sample material, M₂Is a collection of species compositions of single molecules in the molecular composition of the intermediate product，M₃For the second set, card represents the number of elements in the set.

catalyst activityAnd the factor P is the pressure value of the environment where the reaction path is located, and the factor alpha is the pressure influence factor corresponding to the reaction rule corresponding to the reaction path.

The above description is only an example of the present invention, and is not intended to limit the present invention, and it is obvious to those skilled in the art that various modifications and variations can be made in the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A method of simulating a product of a petroleum processing plant, comprising:

determining the molecular composition of the feedstock;

inputting the molecular composition of the raw material into a pre-trained product prediction model; wherein the product prediction model corresponds to a type of petroleum processing plant;

and obtaining the yield of the target product under different reaction conditions output by the product prediction model.

2. The method of claim 1, further comprising, after obtaining yields of target products under different reaction conditions output by the product prediction model:

inquiring the yield of the target product under different reaction conditions;

determining a reaction condition corresponding to the maximum yield value of the target product; and/or the presence of a gas in the gas,

and determining the reaction condition corresponding to the minimum yield value of the target product.

3. The method of claim 2, further comprising, after obtaining yields of target products under different reaction conditions output by the product prediction model:

and determining the physical properties of the target product under different reaction conditions according to the molecular composition of the target product under different reaction conditions.

4. The method of any one of claims 1-3, wherein obtaining yields of target product under different reaction conditions output by the product prediction model comprises:

obtaining the content of each product molecule output by the product prediction model under different reaction conditions;

and obtaining the content of the target product molecules under different reaction conditions from the content of each product molecule under different reaction conditions.

5. The method of any one of claims 1-3, wherein determining the molecular composition of the feedstock comprises:

determining the molecular composition of the feedstock by one or more of comprehensive two-dimensional gas chromatography, quadrupole gas chromatography-mass spectrometry detection, gas chromatography/field ionization-time-of-flight mass spectrometry detection, gas chromatography, near infrared spectroscopy, nuclear magnetic resonance spectroscopy, raman spectroscopy, fourier transform ion cyclotron resonance mass spectrometry, electrostatic field orbitrap mass spectrometry, and ion mobility mass spectrometry.

6. The method of claim 1, wherein the step of training the product prediction model comprises:

establishing a product prediction model; wherein the product prediction model comprises: a reaction rule set and a reaction rate algorithm; the set of reaction rules comprises a plurality of reaction rules;

acquiring sample raw material information of sample raw materials;

training the reaction rule set by using the sample raw material information, and fixing the trained reaction rule set;

and training the reaction rate algorithm by using the sample raw material information, and fixing the trained reaction rate algorithm to obtain the trained product prediction model.

7. The method of claim 6,

sample material information of the sample material includes: the molecular composition of the sample raw material, the molecular content of each molecule in the sample raw material, the second molecular composition of the actual product corresponding to the sample raw material and the actual content of each molecule in the actual product.

8. The method of claim 7, wherein training the set of reaction rules using the sample material information comprises:

processing the molecular composition of the sample raw material according to a preset reaction rule set to obtain a reaction path corresponding to each molecule in the molecular composition of the sample raw material;

obtaining a first molecular composition of a device product according to a reaction path corresponding to each molecule in the molecular composition of the sample raw material; in the device product, comprising: the sample feedstock, intermediate product, and predicted product;

calculating a first relative deviation from a first molecular composition of the device product and a second molecular composition of the actual product;

if the first relative deviation meets a preset condition, fixing the reaction rule set;

and if the first relative deviation does not accord with the preset condition, adjusting the reaction rule in the reaction rule set, and recalculating the first relative difference value according to the adjusted reaction rule set until the first relative deviation accords with the preset condition.

9. The method of claim 8, wherein calculating a first relative deviation from a first molecular composition of the device product and a second molecular composition of the actual product comprises:

acquiring the types of single molecules in the first molecular composition to form a first set;

acquiring the species of single molecules in the second molecular composition to form a second set;

determining whether the second set is a subset of the first set;

if the second set is not the subset of the first set, acquiring a pre-stored relative deviation value which does not meet a preset condition as the first relative deviation value;

if the second set is a subset of the first set, calculating a first relative deviation by:

10. The method of claim 7, wherein training the reaction rate algorithm using the sample feedstock information comprises:

respectively calculating the reaction rate of the reaction path corresponding to each molecule in the molecular composition of the sample raw material according to the reaction rate algorithm;

obtaining the predicted content of each molecule in the predicted product corresponding to the sample raw material according to the molecular content of each molecule in the sample raw material and the reaction rate corresponding to the reaction path of the molecule;

calculating a second relative deviation according to the predicted content of each molecule in the predicted product and the actual content of each molecule in the actual product;

if the second relative deviation meets a preset condition, fixing the reaction rate algorithm;

and if the second relative deviation does not accord with the preset condition, adjusting parameters in the reaction rate algorithm, and recalculating the second relative deviation according to the adjusted reaction rate algorithm until the second relative deviation accords with the preset condition.

11. The method of claim 10, wherein separately calculating the reaction rate of the reaction pathway corresponding to each molecule in the molecular composition of the sample material according to the reaction rate algorithm comprises:

calculating the reaction rate of each reaction path according to the reaction rate constant in the reaction rate algorithm; wherein the reaction rate constant is determined according to the following calculation formula:

12. The method of claim 1, wherein the category of petroleum processing equipment comprises:

catalytic cracker, delayed coking unit, residue hydrogenation unit, hydrocracking unit, diesel oil hydrogenation modification unit, diesel oil hydrogenation refining unit, gasoline hydrogenation refining unit, catalytic reforming unit and alkylation unit; wherein each petroleum processing plant corresponds to a set of reaction rules.

13. The method of claim 1,

the reaction conditions include: reaction temperature, reaction pressure and space velocity.

14. A product simulation system for a petroleum processing plant, the system comprising:

an acquisition unit for determining the molecular composition of the feedstock;

15. The system of claim 14, further comprising:

a second processing unit for querying the yield of the target product under different reaction conditions; determining a reaction condition corresponding to the maximum yield value of the target product; and/or determining the reaction condition corresponding to the minimum yield value of the target product.

16. The system of claim 14, further comprising:

the model training unit is used for establishing a product prediction model; acquiring sample raw material information of sample raw materials; training a reaction rule set included by a product prediction model by using the sample raw material information, and fixing the trained reaction rule set; training a reaction rate algorithm included in a product prediction model by using the sample raw material information, and fixing the reaction rate algorithm after the training to obtain the product prediction model after the training, wherein the product prediction model comprises: a reaction rule set and a reaction rate algorithm; the set of reaction rules includes a plurality of reaction rules.

17. A product simulation apparatus of a petroleum processing plant, characterized in that the product simulation apparatus of the petroleum processing plant comprises a processor, a memory; the processor is configured to execute a product simulation program of a petroleum processing plant stored in the memory to implement the product simulation method of the petroleum processing plant of any one of claims 1 to 13.

18. A storage medium storing one or more programs executable by one or more processors to implement the product simulation method for a petroleum processing plant of any one of claims 1-13.