Chemical looping demercuration composite oxygen carrier optimization design method and device based on machine learning

Abstract

The invention discloses a machine learning-based chemical looping demercuration composite oxygen carrier optimization design method and device. The method comprises the following steps: establishing a composite oxygen carrier macroscopic database; constructing a composite oxygen carrier model; constructing a microscopic source domain data set of the composite oxygen carrier; dividing data in the database into a training set and a test set in proportion; determining the most significant variables of the composite oxygen carrier model influencing the adsorption energy of elemental mercury on the surface of a new composite oxygen carrier, the reaction energy barrier on the surface of the oxygen carrier, the oxygen vacancy formation energy of the new oxygen carrier and the oxygen migration energy barrier, and establishing an optimal feature set; selecting a machine learning algorithm; judging the mercury adsorption and removal capability of the oxygen carrier; and testing and verifying the reliability of the composite oxygen carrier oxygen release performance and demercuration efficiency prediction model by using the test set, and evaluating by using the evaluation index to obtain an optimal composite oxygen carrier oxygen release performance and demercuration efficiency prediction model. According to the method, the experiment trial and error time can be greatly saved, the cost is saved, the efficiency is improved, and a new insight is provided for screening the composite oxygen carrier.

Description

Technical Field

[0001] The present invention belongs to the field of environmental protection technology, and specifically relates to a method and device for optimizing the design of a chemical-chaining demercuration composite oxygen carrier based on machine learning. Background Art

[0002] Coal chemical looping combustion (CLC) is considered to be one of the low-cost CO2 emission reduction technologies with great development and application prospects. It can achieve significant energy savings through its inherent CO2 gas separation characteristics. However, during the coal combustion process, mercury evaporates and is mainly converted into Hg 0 Mercury exists in flue gas in vapor form and, upon emission, spreads through the atmosphere, causing irreversible damage to the ecological environment. Furthermore, mercury in flue gas can react with metals to form amalgams, seriously threatening the operational safety of carbon capture equipment. Therefore, deep flue gas mercury removal in chemical looping systems is imperative.

[0003] Oxygen carriers are the core elements in the chemical looping combustion reaction process. 0 Oxidation can remove elemental mercury, which is difficult to capture, while also enriching CO2 in the flue gas, achieving the dual effects of heavy metal mercury removal and CO2 enrichment. Currently, iron-based oxygen carriers are widely studied due to their good economic benefits. However, their reactivity and mercury removal performance still need to be improved. This can be optimized by adding other elements to construct composite oxygen carriers.

[0004] Nowadays, machine learning and quantum chemistry are increasingly being used for material screening and adsorption energy calculations, improving model calculation performance through statistical methods. Therefore, it is necessary to propose a novel and efficient screening method that combines existing oxygen carrier test data with machine learning to accelerate the development of new oxygen carrier materials, reduce the time and cost of material development, and evaluate material performance. Summary of the Invention

[0005] The object of the present invention is to provide a method and device for optimizing the design of a composite oxygen carrier for chemical looping demercuration based on machine learning, which can reduce the development cost and development cycle of the oxygen carrier and solve the problems described in the background technology.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] The machine learning-based optimization design method for chemical-looping mercury removal composite oxygen carriers includes the following steps:

[0008] Step 1: Collect and organize experimental or simulation data of known composite oxygen carriers and establish a database;

[0009] Step 2: Use quantum chemistry software to construct a composite oxygen carrier model;

[0010] Step 3: Calculate the adsorption energy of elemental mercury on the oxygen carrier surface, the adsorption energy of elemental mercury on the oxygen carrier surface, the reaction energy barrier of elemental mercury on the oxygen carrier surface, the oxygen vacancy formation energy, and the oxygen migration energy barrier based on the composite oxygen carrier model constructed in Step 2 and density functional theory, and construct a source domain data set for the composite oxygen carrier;

[0011] Step 4: Perform preliminary screening and normalization on the experimental or simulated data in step 1, and divide the data in the database into training sets and test sets in proportion;

[0012] Step 5: Preprocess and normalize the source domain dataset from Step 3, and determine the variables that most significantly influence the adsorption energy of elemental mercury on the composite oxygen carrier surface, the reaction energy barrier of elemental mercury on the oxygen carrier surface, the oxygen vacancy formation energy, and the oxygen migration energy barrier of the composite oxygen carrier model through weight analysis to form the optimal feature set;

[0013] Step 6: Select a machine learning algorithm based on the optimal feature set, use the training set to determine the hyperparameters of the machine learning algorithm, and use the machine learning algorithm after the hyperparameters are determined to establish a prediction model for the oxygen release performance and mercury removal efficiency of the composite oxygen carrier based on the training set;

[0014] Step 7: Using quantum chemical calculations to obtain performance indicators of the composite oxygen carrier, the performance of the oxygen carrier is determined to determine its ability to adsorb and remove mercury. This performance indicator is used to evaluate the accuracy of the prediction model for the oxygen release performance and mercury removal efficiency of the composite oxygen carrier. If the accuracy meets the standards, step 8 is performed. If the accuracy does not meet the standards, the machine learning algorithm is replaced and steps 6 and 7 are repeated until the accuracy meets the standards.

[0015] Step 8. Use the regression coefficient R 2 The score and mean absolute error (MAE) were used as evaluation indicators. The reliability of the prediction model for the oxygen release performance and mercury removal efficiency of the composite oxygen carrier was verified using a test set test and evaluated using the evaluation indicators. The optimal prediction model for the oxygen release characteristics and mercury removal efficiency of the composite oxygen carrier was obtained. The oxygen carrier design can be optimized by predicting and calculating the composite oxygen carrier through this model, thereby obtaining a composite oxygen carrier design with ideal oxygen release characteristics and mercury removal efficiency.

[0016] A further improvement of the present invention is that the database in step 1 includes as data the elemental composition of the composite oxygen carrier and the molar mass fraction of each element, the physical properties of the oxygen carrier, the mercury removal characteristics of the oxygen carrier during the chemical looping combustion process, the oxygen release performance and the mercury removal efficiency.

[0017] A further improvement of the present invention is that the composite oxygen carrier model in step 2 is used to calculate the total energy of the model after elemental mercury is adsorbed on the surface of the oxygen carrier, the total energy after elemental mercury optimization, the total energy of the composite oxygen carrier model, the transition state structure energy during the reaction, the energy before and after oxygen vacancy formation, and the energy before and after lattice oxygen migration.

[0018] A further improvement of the present invention is that the adsorption energy of elemental mercury on the surface of the oxygen carrier in step 3 is recorded as ΔE ads , the calculation formula for the accurate value obtained by DFT calculation is as follows:

[0019]

[0020] in, is the system energy of the composite oxygen carrier after adsorption of elemental mercury, eV; E zyt is the system energy of the composite oxygen carrier, eV; is the system energy of elemental mercury, eV;

[0021] The energy barrier of the reaction of elemental mercury on the surface of the oxygen carrier is denoted as ΔE barrier , the calculation formula for the accurate value obtained by DFT calculation is as follows:

[0022]

[0023] Among them, E TS is the transition state structural energy during the post-adsorption reaction, eV; is the system energy of the composite oxygen carrier after adsorption of elemental mercury, eV;

[0024] The oxygen vacancy formation energy is denoted as ΔE f , the calculation formula for the accurate value can be obtained through DFT calculation as follows:

[0025] ΔE f =E T (def)-E T (per)-N O μ O

[0026] Among them, E T (def) is the energy of the oxygen carrier with oxygen vacancies, eV; E T (per) is the energy before the formation of oxygen vacancy of oxygen carrier, eV; N O is the number of oxygen atoms added or reduced (negative when oxygen decreases, positive when oxygen increases), μ O is the chemical potential of oxygen atoms, eV;

[0027] The oxygen migration energy barrier is denoted as ΔE ob , the calculation formula for the accurate value can be obtained through DFT calculation as follows:

[0028] ΔE ob =E TS-O -E IS

[0029] Among them, E TS-Ois the energy of the transition state structure during lattice oxygen migration, eV; E IS is the system energy of the composite oxygen carrier after lattice oxygen migration, eV.

[0030] A further improvement of the present invention is that the source domain data set of the composite oxygen carrier model in step three includes: oxygen carrier structural model information, characteristic variables corresponding to the structural model of the oxygen carrier, adsorption energy of elemental mercury on the surface of the composite oxygen carrier, reaction energy barrier of elemental mercury on the surface of the oxygen carrier, oxygen vacancy formation energy and oxygen migration energy barrier.

[0031] A further improvement of the present invention is that the composite oxygen carrier model information in step three includes the basic configuration of the composite oxygen carrier, active site information, doping elements, and doping type information; and the characteristic variables refer to atomic radius, partial wave state density, charge transfer amount, Mulliken charge distribution and differential charge density.

[0032] A further improvement of the present invention is that in step 6, the machine learning algorithms used include: support vector machine, generalized regression neural network and random forest learner regression method;

[0033] The elemental composition, physical properties and experimental conditions of the composite oxygen carrier were determined as the input parameters in the training set; the experimental data of the composite oxygen carrier's oxygen release performance and mercury removal efficiency were used as output parameters. The oxygen release performance and mercury removal efficiency prediction model of the composite oxygen carrier was used to calculate the oxygen release performance and mercury removal efficiency of the composite oxygen carrier under the same experimental conditions.

[0034] A further improvement of the present invention is that, in the performance index of the composite oxygen carrier in step seven, the more negative the adsorption energy of elemental mercury on the surface of the new composite oxygen carrier, the stronger the interaction between elemental mercury and the surface of the oxygen carrier, the more stable the adsorption system, and the stronger the performance of the oxygen carrier;

[0035] Among the performance indicators of composite oxygen carriers, the reaction energy barrier of elemental mercury on the surface of the oxygen carrier is the minimum energy required for the reaction to occur. The lower the reaction energy barrier, the easier it is for the reaction to occur, indicating that the performance of the oxygen carrier is stronger.

[0036] Among the performance indicators of composite oxygen carriers, the smaller the oxygen vacancy formation energy of the composite oxygen carrier, the easier it is to form oxygen defects on the surface of the oxygen carrier, and the stronger the performance of the oxygen carrier in generating active oxygen;

[0037] Among the performance indicators of composite oxygen carriers, the oxygen migration energy barrier is the minimum energy required for the lattice oxygen migration reaction of the composite oxygen carrier to occur. If the oxygen migration energy barrier is lower, the reaction is easier to occur, indicating that the migration ability of the lattice oxygen inside the oxygen carrier is stronger.

[0038] A further improvement of the present invention is that in step eight, the regression coefficient R is used 2The score and mean absolute error MAE are used as evaluation indicators, and the calculation method is as follows:

[0039]

[0040] Where N is the total number of samples, Y i is the true value, y i is the adsorption energy predicted by the machine learning algorithm, is the average of all values of the true value;

[0041]

[0042] Where N is the total number of samples, Y i is the true value, y i is the adsorption energy predicted by the machine learning algorithm, is the average of all true values.

[0043] The machine learning-based chemical looping demercuration composite oxygen carrier optimization design device includes:

[0044] Data collection and collation unit, collects and organizes experimental or simulation data of known composite oxygen carriers and establishes a database;

[0045] Composite oxygen carrier model building unit, using quantum chemistry software to build composite oxygen carrier model;

[0046] A source domain data set construction unit calculates the adsorption energy of elemental mercury on the surface of the oxygen carrier, the adsorption energy of elemental mercury on the surface of the oxygen carrier, the reaction energy barrier of elemental mercury on the surface of the oxygen carrier, the oxygen vacancy formation energy and the oxygen migration energy barrier based on the composite oxygen carrier model constructed by the model construction unit and density functional theory, and constructs a source domain data set of the composite oxygen carrier;

[0047] The data processing unit performs preliminary screening and normalization on the experimental or simulation data in the data collection and collation unit, and divides the data in the database into training sets and test sets in proportion;

[0048] The optimal feature set forming unit preprocesses and normalizes the source domain data set in the source domain data set building unit, and determines the variables that have the most significant influence on the adsorption energy of elemental mercury on the composite oxygen carrier surface, the reaction energy barrier of elemental mercury on the oxygen carrier surface, the oxygen vacancy formation energy, and the oxygen migration energy barrier of the composite oxygen carrier model through weight analysis, thereby forming the optimal feature set;

[0049] A prediction model building unit selects a machine learning algorithm based on the optimal feature set, determines hyperparameters of the machine learning algorithm using the training set, and uses the machine learning algorithm after determining the hyperparameters to establish a prediction model for the oxygen release performance and mercury removal efficiency of the composite oxygen carrier based on the training set;

[0050] The prediction model training unit uses quantum chemical calculations to obtain the performance indicators of the composite oxygen carrier, which are used to determine the oxygen carrier's ability to adsorb and remove mercury. This is used as an evaluation indicator to evaluate the accuracy of the prediction model for the composite oxygen carrier's oxygen release performance and mercury removal efficiency. If the accuracy meets the standard, the evaluation unit is executed. If the accuracy does not meet the standard, the machine learning algorithm is replaced and the optimal feature set formation unit and the prediction model construction unit are repeated until the accuracy meets the standard.

[0051] Evaluation unit, using regression coefficient R 2 The score and mean absolute error (MAE) were used as evaluation indicators. The reliability of the prediction model for the oxygen release performance and mercury removal efficiency of the composite oxygen carrier was verified using a test set test and evaluated using the evaluation indicators. The optimal prediction model for the oxygen release characteristics and mercury removal efficiency of the composite oxygen carrier was obtained. The oxygen carrier design can be optimized by predicting and calculating the composite oxygen carrier through this model, thereby obtaining a composite oxygen carrier design with ideal oxygen release characteristics and mercury removal efficiency.

[0052] Compared with the prior art, the present invention has at least the following beneficial technical effects:

[0053] The present invention provides a method and device for optimizing the design of composite oxygen carriers for chemical-linking mercury removal based on machine learning. Based on information such as the elemental composition and molar mass fraction of each element of a known composite oxygen carrier, as well as information such as the physical properties (specific surface area) of the oxygen carrier, machine learning is used to predict the oxygen-release characteristics and mercury removal efficiency of an unknown composite oxygen carrier. Traditional trial-and-error methods require a large number of synthesis and testing experiments. The present invention can significantly save experimental trial-and-error time, save costs, and improve efficiency, providing new insights into the screening of composite oxygen carriers. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0055] Figure 1 It is a schematic flow chart of the method of the present invention;

[0056] Figure 2 This is a diagram of the oxygen carrier performance test experimental system;

[0057] Figure 3 This is a molecular model diagram of Cl-modified copper-iron composite oxygen carrier;

[0058] Figure 4 This is a molecular model diagram of Hg adsorbed on the surface O atoms of Cl-modified copper-iron composite oxygen carrier;

[0059] Figure 5 This is the calculated result of the transition state of the oxidation reaction of elemental mercury after adsorption on Cl-modified copper-iron composite oxygen carrier;

[0060] Figure 6 It is the partial wave state density result when Hg is adsorbed on the surface O atoms of Cl-modified copper-iron composite oxygen carrier;

[0061] Figure 7 This is a structural block diagram of the device for optimizing the design of chemical-chaining mercury removal composite oxygen carriers based on machine learning. DETAILED DESCRIPTION

[0062] Hereinafter, only certain exemplary embodiments are briefly described. As will be appreciated by those skilled in the art, the described embodiments may be modified in various ways without departing from the spirit or scope of the present invention. Therefore, the drawings and description are to be considered as illustrative in nature and not restrictive.

[0063] In the description of the present invention, it should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or collections thereof.

[0064] It should also be understood that the terms used in the present specification are only for the purpose of describing particular embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, the singular forms "a", "an", and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0065] It should be further understood that the term "and / or" used in the present description and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0066] The accompanying drawings illustrate various schematic diagrams of structures according to embodiments disclosed herein. These figures are not drawn to scale; for clarity, some details are exaggerated and some details may be omitted. The shapes of the various regions and layers shown in the figures, as well as their relative sizes and positional relationships, are merely exemplary and may deviate in practice due to manufacturing tolerances or technical limitations. Those skilled in the art may design regions / layers with different shapes, sizes, and relative positions as needed.

[0067] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0068] Example 1

[0069] like Figure 1 As shown, the present invention provides a method for optimizing the design of a composite oxygen carrier for chemical chaining mercury removal based on machine learning, comprising:

[0070] Step 1: Collect and organize experimental or simulation data of known composite oxygen carriers and establish a database;

[0071] The database includes information such as the element composition of the composite oxygen carrier and the molar mass fraction of each element, information such as the physical properties of the oxygen carrier (specific surface area, etc.), information such as the oxygen release performance and mercury removal efficiency of the oxygen carrier during the chemical looping combustion process, wherein the oxygen release performance and mercury removal efficiency of the oxygen carrier are obtained by Figure 2 The oxygen carrier performance test experimental system shown is obtained.

[0072] Step 2: Use quantum chemistry software to construct a composite oxygen carrier model;

[0073] The quantum chemistry software, such as Materials Studio, VASP, etc.;

[0074] Based on the characteristics of atomic radius, partial wave state density, charge transfer amount, Mulliken charge population and differential charge density, composite oxygen carriers with oxygen release and mercury removal properties are selected, and the basic configuration, active site information, doping elements, and doping positions of these oxygen carriers are collected to establish a structural model of the composite oxygen carrier, such as Figure 3 The molecular model diagram of Cl-modified copper-iron composite oxygen carrier;

[0075] The composite oxygen carrier model is used to calculate the total energy of the model after elemental mercury is adsorbed on the surface of the oxygen carrier, the total energy after elemental mercury optimization, the total energy of the composite oxygen carrier model, the energy of the transition state structure during the reaction, the energy before and after the formation of oxygen vacancies, and the energy before and after the migration of lattice oxygen;

[0076] Step 3: Calculate the adsorption energy of elemental mercury on the oxygen carrier surface, the adsorption energy of elemental mercury on the oxygen carrier surface, the reaction energy barrier of elemental mercury on the oxygen carrier surface, the oxygen vacancy formation energy, and the oxygen migration energy barrier based on the composite oxygen carrier model constructed in Step 2 and density functional theory, and construct a source domain data set for the composite oxygen carrier;

[0077] The specific steps for step three are as follows:

[0078] Substitute the established composite oxygen carrier model into the composite oxygen carrier model established in step 2 to obtain the total energy of the model after elemental mercury adsorption on the oxygen carrier surface, the total energy after elemental mercury optimization, the total energy of the composite oxygen carrier model, the transition state structure energy during the reaction, the energy before and after oxygen vacancy formation, and the energy before and after lattice oxygen migration, for example Figure 4This is a molecular model diagram of Hg adsorbed on the surface O atoms of Cl-modified copper-iron composite oxygen carrier, which is used to calculate the adsorption energy of elemental mercury. Figure 5 This is the calculated result of the transition state of the oxidation reaction after elemental mercury adsorption;

[0079] The adsorption energy of elemental mercury on the surface of the oxygen carrier is recorded as ΔE ads , the calculation formula for the accurate value can be obtained through DFT calculation as follows:

[0080]

[0081] in, is the system energy of the composite oxygen carrier after adsorption of elemental mercury, eV; E zyt is the system energy of the composite oxygen carrier, eV; is the system energy of elemental mercury, eV.

[0082] The energy barrier of the reaction of elemental mercury on the surface of the oxygen carrier is denoted as ΔE barrier , the calculation formula for the accurate value can be obtained through DFT calculation as follows:

[0083]

[0084] Among them, E TS is the transition state structural energy during the post-adsorption reaction, eV; is the system energy of the composite oxygen carrier after adsorption of elemental mercury, eV.

[0085] The oxygen vacancy formation energy in the composite oxygen carrier is recorded as ΔE f , the calculation formula for the accurate value can be obtained through DFT calculation as follows:

[0086] ΔE f =E T (def)-E T (per)-N O μ O

[0087] Among them, E T (def) is the energy of the oxygen carrier with oxygen vacancies, eV; E T (per) is the energy before the formation of oxygen vacancy of oxygen carrier, eV; N O is the number of oxygen atoms added or reduced (negative when oxygen decreases, positive when oxygen increases), μ O is the chemical potential of the oxygen atom, eV.

[0088] The energy barrier of lattice oxygen migration in the composite oxygen carrier is denoted as ΔE ob , the calculation formula for the accurate value can be obtained through DFT calculation as follows:

[0089] ΔEob =E TS-O -E IS

[0090] Among them, E TS-O is the energy of the transition state structure during lattice oxygen migration, eV; E IS is the system energy of the composite oxygen carrier after lattice oxygen migration, eV.

[0091] Furthermore, a source domain dataset of the composite oxygen carrier model is established;

[0092] The source domain data set includes information on the oxygen carrier structure model, characteristic variables corresponding to the oxygen carrier structure model, adsorption energy of elemental mercury on the composite oxygen carrier surface, reaction energy barrier of elemental mercury on the oxygen carrier surface, oxygen vacancy formation energy, and oxygen migration energy barrier;

[0093] The composite oxygen carrier model information includes the basic configuration of the composite oxygen carrier, active site information, doping elements, and doping type information; the characteristic variables refer to atomic radius, partial wave state density, charge transfer amount, Mulliken charge distribution and differential charge density, etc., for example Figure 6 The partial wave density of states results when Hg is adsorbed on the surface O atoms of Cl-modified copper-iron composite oxygen carrier;

[0094] Step 4: Perform preliminary screening and normalization on the experimental or simulated data in step 1, and divide the data in the database into training sets and test sets in proportion;

[0095] The training and test sets were randomly divided into an 8:2 ratio, with the training set accounting for 80% and the test set accounting for 20%. The training set was used to train the composite oxygen carrier prediction model, while the test set was used to test and verify the accuracy of the model and was used after the model training was completed.

[0096] Step 5: Preprocess and normalize the source domain dataset from Step 3, and determine the variables that most significantly influence the adsorption energy of elemental mercury on the composite oxygen carrier surface, the reaction energy barrier of elemental mercury on the oxygen carrier surface, the oxygen vacancy formation energy, and the oxygen migration energy barrier of the composite oxygen carrier model through weight analysis to form the optimal feature set;

[0097] Step 6: Select a machine learning algorithm, use the training set to determine the hyperparameters of the machine learning algorithm, and use the machine learning algorithm after determining the hyperparameters to establish a prediction model for the oxygen release performance and mercury removal efficiency of the composite oxygen carrier based on the training set.

[0098] In step 6, the machine learning algorithms used include but are not limited to: support vector machine (SVM), generalized regression neural network (GRNN), random forest (RF) and other learner regression methods;

[0099] In this work, parameters such as the elemental composition, physical properties (such as specific surface area), and experimental conditions (such as reaction temperature) of the composite oxygen carrier were selected as input parameters in the training set. Experimental data on the composite oxygen carrier's oxygen release performance and mercury removal efficiency were used as output parameters. A prediction model for composite oxygen carrier oxygen release performance and mercury removal efficiency was used to calculate the composite oxygen carrier's oxygen release performance and mercury removal efficiency under the same experimental conditions.

[0100] Step 7: Using quantum chemical calculations to obtain performance indicators of the composite oxygen carrier, the performance of the composite oxygen carrier is determined to determine its ability to adsorb and remove mercury. This performance indicator is used to evaluate the accuracy of the prediction model for the composite oxygen carrier's oxygen release performance and mercury removal efficiency. If the accuracy meets the standards, step 8 is performed. If the accuracy does not meet the standards, the machine learning algorithm is replaced and steps 6 and 7 are repeated until the accuracy meets the standards.

[0101] The performance indicators of the composite oxygen carrier described in step seven include the adsorption energy of elemental mercury on the surface of the new composite oxygen carrier, the reaction energy barrier of elemental mercury on the surface of the oxygen carrier, the oxygen vacancy formation energy and the oxygen migration energy barrier, the oxygen release performance of the oxygen carrier during the chemical looping combustion process, the mercury removal efficiency and other information.

[0102] Furthermore, if the adsorption energy of elemental mercury on the surface of the new composite oxygen carrier is more negative, it means that the interaction between elemental mercury and the surface of the oxygen carrier is stronger, the adsorption system is more stable, and the mercury adsorption performance of the oxygen carrier is stronger.

[0103] Furthermore, the reaction energy barrier of elemental mercury on the surface of the oxygen carrier is the minimum energy required for the reaction to occur. If the reaction energy barrier is lower, the reaction is easier to occur, indicating that the mercury oxidation performance of the oxygen carrier is stronger.

[0104] Furthermore, if the oxygen vacancy formation energy of the composite oxygen carrier is smaller, it means that it is easier for oxygen defects to form on the surface of the oxygen carrier, and the performance of the oxygen carrier in generating active oxygen is stronger.

[0105] Furthermore, the oxygen migration energy barrier of the composite oxygen carrier is the minimum energy required for the lattice oxygen migration reaction of the composite oxygen carrier to occur. If the oxygen migration energy barrier is lower, the reaction is easier to occur, indicating that the migration ability of the lattice oxygen inside the oxygen carrier is stronger.

[0106] Step 8: Using a test set to test and verify the reliability of the composite oxygen carrier oxygen release characteristics and mercury removal efficiency prediction model and evaluate it with evaluation indicators to obtain the optimal composite oxygen carrier oxygen release performance and mercury removal efficiency prediction model;

[0107] In step 8, the regression coefficient R is used 2 The score and mean absolute error MAE are used as evaluation indicators, and the calculation method is as follows:

[0108]

[0109] Where N is the total number of samples, Yi is the true value, y i is the adsorption energy predicted by the machine learning algorithm, is the average of all values of the true value;

[0110]

[0111] Where N is the total number of samples, Y i is the true value, y i is the adsorption energy predicted by the machine learning algorithm, is the average of all values of the true value;

[0112] According to steps 1 to 8 of the present invention, a composite oxygen carrier with the best oxygen release performance and mercury removal efficiency index is screened out.

[0113] Example 2

[0114] like Figure 7 As shown, the device for optimizing the design of composite oxygen carriers for chemical looping mercury removal based on machine learning provided by the present invention includes:

[0115] Data collection and collation unit, collects and organizes experimental or simulation data of known composite oxygen carriers and establishes a database;

[0116] Composite oxygen carrier model building unit, using quantum chemistry software to build composite oxygen carrier model;

[0117] A source domain data set construction unit calculates the adsorption energy of elemental mercury on the surface of the oxygen carrier, the adsorption energy of elemental mercury on the surface of the oxygen carrier, the reaction energy barrier of elemental mercury on the surface of the oxygen carrier, the oxygen vacancy formation energy and the oxygen migration energy barrier based on the composite oxygen carrier model constructed by the model construction unit and density functional theory, and constructs a source domain data set of the composite oxygen carrier;

[0118] The data processing unit performs preliminary screening and normalization on the experimental or simulation data in the data collection and collation unit, and divides the data in the database into training sets and test sets in proportion;

[0119] The optimal feature set forming unit preprocesses and normalizes the source domain data set in the source domain data set building unit, and determines the variables that have the most significant influence on the adsorption energy of elemental mercury on the composite oxygen carrier surface, the reaction energy barrier of elemental mercury on the oxygen carrier surface, the oxygen vacancy formation energy, and the oxygen migration energy barrier of the composite oxygen carrier model through weight analysis, thereby forming the optimal feature set;

[0120] A prediction model building unit selects a machine learning algorithm based on the optimal feature set, determines hyperparameters of the machine learning algorithm using the training set, and uses the machine learning algorithm after determining the hyperparameters to establish a prediction model for the oxygen release performance and mercury removal efficiency of the composite oxygen carrier based on the training set;

[0121] The prediction model training unit uses quantum chemical calculations to obtain the performance indicators of the composite oxygen carrier, which are used to determine the oxygen carrier's ability to adsorb and remove mercury. This is used as an evaluation indicator to evaluate the accuracy of the prediction model for the composite oxygen carrier's oxygen release performance and mercury removal efficiency. If the accuracy meets the standard, the evaluation unit is executed. If the accuracy does not meet the standard, the machine learning algorithm is replaced and the optimal feature set formation unit and the prediction model construction unit are repeated until the accuracy meets the standard.

[0122] Evaluation unit, using regression coefficient R 2 The score and mean absolute error (MAE) were used as evaluation indicators. The reliability of the prediction model for the oxygen release performance and mercury removal efficiency of the composite oxygen carrier was verified using a test set test and evaluated using the evaluation indicators. The optimal prediction model for the oxygen release characteristics and mercury removal efficiency of the composite oxygen carrier was obtained. The oxygen carrier design can be optimized by predicting and calculating the composite oxygen carrier through this model, thereby obtaining a composite oxygen carrier design with ideal oxygen release characteristics and mercury removal efficiency.

[0123] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention. Therefore, from all points of view, the embodiments should be regarded as illustrative and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description, and it is intended that all changes that fall within the meaning and range of equivalents of the claims are included in the present invention. Any reference signs in the claims should not be construed as limiting the claim to which they relate.

[0124] In addition, it should be understood that although this specification describes the embodiments, not every embodiment contains only one independent technical solution. This description is for clarity only. Those skilled in the art should consider the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art. The above content is only for the purpose of illustrating the technical concept of the present invention and cannot be used to limit the scope of protection of the present invention. Any changes made based on the technical solution in accordance with the technical concept proposed by the present invention fall within the scope of protection of the claims of the present invention.

Claims

1. A method for optimizing the design of composite oxygen carriers for chemical chaining mercury removal based on machine learning, characterized in that: The following steps are involved: Step 1: Collect and organize experimental or simulation data of known composite oxygen carriers and establish a database; Step 2: Use quantum chemistry software to construct a composite oxygen carrier model; Step 3: Calculate the adsorption energy of elemental mercury on the oxygen carrier surface, the adsorption energy of elemental mercury on the oxygen carrier surface, the reaction energy barrier of elemental mercury on the oxygen carrier surface, the oxygen vacancy formation energy, and the oxygen migration energy barrier based on the composite oxygen carrier model constructed in Step 2 and density functional theory, and construct a source domain data set for the composite oxygen carrier; Step 4: Perform preliminary screening and normalization on the experimental or simulated data in step 1, and divide the data in the database into training sets and test sets in proportion; Step 5: Preprocess and normalize the source domain dataset from Step 3, and determine the variables that most significantly influence the adsorption energy of elemental mercury on the composite oxygen carrier surface, the reaction energy barrier of elemental mercury on the oxygen carrier surface, the oxygen vacancy formation energy, and the oxygen migration energy barrier of the composite oxygen carrier model through weight analysis to form the optimal feature set; Step 6: Select a machine learning algorithm based on the optimal feature set, use the training set to determine the hyperparameters of the machine learning algorithm, and use the machine learning algorithm after the hyperparameters are determined to establish a prediction model for the oxygen release performance and mercury removal efficiency of the composite oxygen carrier based on the training set; Step 7: Using quantum chemical calculations to obtain performance indicators of the composite oxygen carrier, the performance of the oxygen carrier is determined to determine its ability to adsorb and remove mercury. This performance indicator is used to evaluate the accuracy of the prediction model for the oxygen release performance and mercury removal efficiency of the composite oxygen carrier. If the accuracy meets the standards, step 8 is performed. If the accuracy does not meet the standards, the machine learning algorithm is replaced and steps 6 and 7 are repeated until the accuracy meets the standards. Step 8. Use the regression coefficient R 2 The score and mean absolute error (MAE) were used as evaluation indicators. The reliability of the prediction model for the oxygen release performance and mercury removal efficiency of the composite oxygen carrier was verified using a test set test and evaluated using the evaluation indicators. The optimal prediction model for the oxygen release characteristics and mercury removal efficiency of the composite oxygen carrier was obtained. The oxygen carrier design can be optimized by predicting and calculating the composite oxygen carrier through this model, thereby obtaining a composite oxygen carrier design with ideal oxygen release characteristics and mercury removal efficiency.

2. The method for optimizing the design of composite oxygen carriers for chemical chaining mercury removal based on machine learning according to claim 1, characterized in that: The database in step 1 includes data on the elemental composition of the composite oxygen carrier and the molar mass fraction of each element, the physical properties of the oxygen carrier, the mercury removal characteristics of the oxygen carrier during chemical looping combustion, oxygen release performance, and mercury removal efficiency.

3. The method for optimizing the design of composite oxygen carriers for chemical chaining mercury removal based on machine learning according to claim 1, characterized in that: The composite oxygen carrier model in step 2 is used to calculate the total energy of the model after elemental mercury is adsorbed on the oxygen carrier surface, the total energy after elemental mercury optimization, the total energy of the composite oxygen carrier model, the transition state structure energy during the reaction, the energy before and after oxygen vacancy formation, and the energy before and after lattice oxygen migration.

4. The method for optimizing the design of composite oxygen carriers for chemical chaining mercury removal based on machine learning according to claim 1, characterized in that: The adsorption energy of elemental mercury on the surface of the oxygen carrier in step 3 is recorded as ΔE ads , the calculation formula for the accurate value obtained by DFT calculation is as follows: in, is the system energy of the composite oxygen carrier after adsorption of elemental mercury, eV; E zyt is the system energy of the composite oxygen carrier, eV; is the system energy of elemental mercury, eV; The energy barrier of the reaction of elemental mercury on the surface of the oxygen carrier is denoted as ΔE barrier , the calculation formula for the accurate value obtained by DFT calculation is as follows: Among them, E TS is the transition state structural energy during the post-adsorption reaction, eV; is the system energy of the composite oxygen carrier after adsorption of elemental mercury, eV; The oxygen vacancy formation energy is denoted as ΔE f , the calculation formula for the accurate value can be obtained through DFT calculation as follows: DE f =E T (def)-E T (per)-N O m O Among them, E T (def) is the energy of the oxygen carrier with oxygen vacancies, eV; E T (per) is the energy before the formation of oxygen vacancy of oxygen carrier, eV; N O is the number of oxygen atoms added or reduced (negative when oxygen decreases, positive when oxygen increases), μ O is the chemical potential of oxygen atoms, eV; The oxygen migration energy barrier is denoted as ΔE ob , the calculation formula for the accurate value can be obtained through DFT calculation as follows: ΔE ob =E TS-O -E IS Among them, E TS-O is the energy of the transition state structure during lattice oxygen migration, eV; E IS is the system energy of the composite oxygen carrier after lattice oxygen migration, eV.

5. The method for optimizing the design of composite oxygen carriers for chemical chaining mercury removal based on machine learning according to claim 1, characterized in that: The source domain data set of the composite oxygen carrier model in step three includes: oxygen carrier structural model information, characteristic variables corresponding to the structural model of the oxygen carrier, adsorption energy of elemental mercury on the composite oxygen carrier surface, reaction energy barrier of elemental mercury on the oxygen carrier surface, oxygen vacancy formation energy and oxygen migration energy barrier.

6. The method for optimizing the design of composite oxygen carriers for chemical chaining mercury removal based on machine learning according to claim 1, characterized in that: The composite oxygen carrier model information in step three includes the basic configuration of the composite oxygen carrier, active site information, doping elements, and doping type information; the characteristic variables refer to atomic radius, partial wave state density, charge transfer amount, Mulliken charge distribution and differential charge density.

7. The method for optimizing the design of composite oxygen carriers for chemical chaining mercury removal based on machine learning according to claim 1, characterized in that: In step 6, the machine learning algorithms used include support vector machine, generalized regression neural network and random forest learner regression method; The elemental composition, physical properties and experimental conditions of the composite oxygen carrier were determined as the input parameters in the training set; the experimental data of the composite oxygen carrier's oxygen release performance and mercury removal efficiency were used as output parameters. The oxygen release performance and mercury removal efficiency prediction model of the composite oxygen carrier was used to calculate the oxygen release performance and mercury removal efficiency of the composite oxygen carrier under the same experimental conditions.

8. The method for optimizing the design of composite oxygen carriers for chemical chaining mercury removal based on machine learning according to claim 1, characterized in that: In the performance index of the composite oxygen carrier in step 7, if the adsorption energy of elemental mercury on the surface of the new composite oxygen carrier is more negative, it means that the interaction between elemental mercury and the surface of the oxygen carrier is stronger, the adsorption system is more stable, and the performance of the oxygen carrier is stronger; Among the performance indicators of composite oxygen carriers, the reaction energy barrier of elemental mercury on the surface of the oxygen carrier is the minimum energy required for the reaction to occur. The lower the reaction energy barrier, the easier it is for the reaction to occur, indicating that the performance of the oxygen carrier is stronger. Among the performance indicators of composite oxygen carriers, the smaller the oxygen vacancy formation energy of the composite oxygen carrier, the easier it is to form oxygen defects on the surface of the oxygen carrier, and the stronger the performance of the oxygen carrier in generating active oxygen; Among the performance indicators of composite oxygen carriers, the oxygen migration energy barrier is the minimum energy required for the lattice oxygen migration reaction of the composite oxygen carrier to occur. If the oxygen migration energy barrier is lower, the reaction is easier to occur, indicating that the migration ability of the lattice oxygen inside the oxygen carrier is stronger.

9. The method for optimizing the design of composite oxygen carriers for chemical chaining mercury removal based on machine learning according to claim 1, characterized in that: In step eight, the regression coefficient R is used 2 The score and mean absolute error MAE are used as evaluation indicators, and the calculation method is as follows: Where N is the total number of samples, Y i is the true value, y i is the adsorption energy predicted by the machine learning algorithm, is the average of all values of the true value; Where N is the total number of samples, Y i is the true value, y i is the adsorption energy predicted by the machine learning algorithm, is the average of all true values.

10. A chemical chaining demercuration composite oxygen carrier optimization design device based on machine learning, characterized in that: include: Data collection and collation unit, collects and organizes experimental or simulation data of known composite oxygen carriers and establishes a database; Composite oxygen carrier model building unit, using quantum chemistry software to build composite oxygen carrier model; A source domain data set construction unit calculates the adsorption energy of elemental mercury on the surface of the oxygen carrier, the adsorption energy of elemental mercury on the surface of the oxygen carrier, the reaction energy barrier of elemental mercury on the surface of the oxygen carrier, the oxygen vacancy formation energy and the oxygen migration energy barrier based on the composite oxygen carrier model constructed by the model construction unit and density functional theory, and constructs a source domain data set of the composite oxygen carrier; The data processing unit performs preliminary screening and normalization on the experimental or simulation data in the data collection and collation unit, and divides the data in the database into training sets and test sets in proportion; The optimal feature set forming unit preprocesses and normalizes the source domain data set in the source domain data set building unit, and determines the variables that have the most significant influence on the adsorption energy of elemental mercury on the composite oxygen carrier surface, the reaction energy barrier of elemental mercury on the oxygen carrier surface, the oxygen vacancy formation energy, and the oxygen migration energy barrier of the composite oxygen carrier model through weight analysis, thereby forming the optimal feature set; A prediction model building unit selects a machine learning algorithm based on the optimal feature set, determines hyperparameters of the machine learning algorithm using the training set, and uses the machine learning algorithm after determining the hyperparameters to establish a prediction model for the oxygen release performance and mercury removal efficiency of the composite oxygen carrier based on the training set; The prediction model training unit uses quantum chemical calculations to obtain the performance indicators of the composite oxygen carrier, which are used to determine the oxygen carrier's ability to adsorb and remove mercury. This is used as an evaluation indicator to evaluate the accuracy of the prediction model for the composite oxygen carrier's oxygen release performance and mercury removal efficiency. If the accuracy meets the standard, the evaluation unit is executed. If the accuracy does not meet the standard, the machine learning algorithm is replaced and the optimal feature set formation unit and the prediction model construction unit are repeated until the accuracy meets the standard. Evaluation unit, using regression coefficient R 2 The score and mean absolute error (MAE) were used as evaluation indicators. The reliability of the prediction model for the oxygen release performance and mercury removal efficiency of the composite oxygen carrier was verified using a test set test and evaluated using the evaluation indicators. The optimal prediction model for the oxygen release characteristics and mercury removal efficiency of the composite oxygen carrier was obtained. The oxygen carrier design can be optimized by predicting and calculating the composite oxygen carrier through this model, thereby obtaining a composite oxygen carrier design with ideal oxygen release characteristics and mercury removal efficiency.

Images