Method and equipment for generating heuristic multi-tooth adsorption configuration based on particle swarm optimization

Abstract

The invention relates to a heuristic multi-tooth adsorption configuration generation method and equipment based on particle swarm optimization, and aims to improve the limitation that a reasonable initial configuration is difficult to generate on a complex surface by an existing automatic generation adsorption configuration algorithm. The algorithm effectively solves the problem of multi-tooth adsorption of the complex adsorbate, and fully considers multiple adsorption configurations of the adsorbate, so that the generated initial configuration is closer to the optimal adsorption configuration. According to the algorithm, the adsorption behavior is deeply understood from the molecular level, and a coordinate layout enabling a molecular adsorption configuration to be highly similar to a parent molecular structure of the molecular adsorption configuration is found through a PSO (Particle Swarm Optimization) process. According to the algorithm, the inherent adsorption preference of adsorbate and the chemical characteristics of the material surface are fully utilized, a reasonable adsorption configuration is rapidly generated, the efficiency and rationality of adsorption structure initialization are remarkably improved, and the algorithm is suitable for automatic modeling of a complex catalytic system.

Description

Technical Field

[0001] The present invention relates to the technical field of computer chemistry, and in particular to a method and device for generating a heuristic multi-dentate adsorption configuration based on particle swarm optimization. Background Art

[0002] In the field of computational chemistry, with the rapid improvement of scientific computing capabilities, the research paradigm has gradually transitioned from the traditional experiment-driven to the data-driven model. Automated modeling technology has played a vital role in this transformation. Computational chemistry usually involves complex molecular systems and chemical reaction processes, requiring a large amount of accurate data and calculations to predict the properties of substances, reaction mechanisms and molecular behaviors. However, existing automated algorithms have certain limitations when dealing with complex systems, especially the multi-dentate adsorption process of adsorbate molecules on complex catalyst surfaces. There is still a problem of difficulty in quickly generating a reasonable initial structure. Existing adsorption structure generation algorithms are usually only for simple adsorption systems, and it is difficult to solve the adsorption behavior of complex surfaces, resulting in unreasonable and inefficient generated adsorption structures.

[0003] Therefore, in the context of the rapid development of high-performance computers and big data, inventing an efficient algorithm to quickly generate reasonable adsorption configurations has become an indispensable tool in the field of computational chemistry. Summary of the invention

[0004] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide a heuristic multi-tooth adsorption configuration generation method and equipment based on particle swarm optimization, so as to solve or partially solve the problem of unreasonable adsorption structure and low efficiency generated under the adsorption behavior of complex surfaces.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] One aspect of the present invention provides a heuristic multi-tooth adsorption configuration generation method based on particle swarm optimization, comprising the following steps:

[0007] Obtain topological structure information of adsorbate molecules and atomic coordinate information of catalyst materials;

[0008] For the adsorbate molecule, the parent molecule is matched based on the topological structure information, and the bond-missing position in the parent molecule is used as the bonding guide vector of the adsorbate molecule, wherein the topological graph of the adsorbate molecule is a subgraph of the topological graph of the parent molecule;

[0009] For the catalyst material, calculating the adsorption vectors of the surface atoms based on the atomic coordinate information;

[0010] Establish a one-to-one mapping relationship between the rigid body model parameter group of the adsorbate molecule and the atomic coordinates within the adsorbate molecule;

[0011] Based on the bonding orientation vector and the adsorption vector, searching for a rigid body model parameter group of an optimal adsorption configuration of adsorbate molecules on the catalyst surface by particle swarm optimization under the premise of considering at least one of structural similarity, adsorption site preference, and micro-environmental influence;

[0012] Based on the rigid body model parameter group of the optimal adsorption configuration, the optimal adsorption configuration of the catalyst surface is restored using the mapping relationship.

[0013] As a preferred technical solution, the process of matching the parent molecule based on the topological structure information and using the missing bond position in the parent molecule as the bonding guide vector of the adsorbate molecule includes the following steps:

[0014] Convert the adsorbate molecule into a topological structure and match one or more parent molecules. After topological matching, the number of missing bonds of the adsorbate molecule is less than or equal to 3, and all missing bonds face the same side.

[0015] The relative positions of atoms in the adsorbate molecule are extracted from the parent molecule, and the missing bond position is used as the bonding guide vector of the adsorbate molecule, wherein the bonding guide vector includes the bonding characteristic information and geometric reference information of the adsorbate molecule in the parent molecule.

[0016] As a preferred technical solution, the process of calculating the adsorption vector of the surface atoms based on the atomic coordinate information includes the following steps:

[0017] Identify the locations of surface atoms of catalyst materials;

[0018] For each surface atom, neighboring atoms are searched within the preset stage radius, the relative position vector and distance between neighboring atoms are calculated, the relative position vector is weighted according to the distance to obtain the adsorption vector of the surface atom and normalized.

[0019] As a preferred technical solution, the process of constructing the mapping relationship includes the following steps:

[0020] Modeling the adsorbate molecule as a six-degree-of-freedom rigid body model determined by a parameter set, wherein the parameter set includes a translation parameter and a rotation parameter;

[0021] Through the special Euclidean group SE(3) transformation, a one-to-one mapping relationship between the rigid body model parameter group and the atomic coordinates of the adsorbate molecules is established, and the spatial configuration of the adsorbate molecules is uniquely determined by the parameter group.

[0022] As a preferred technical solution, in the particle swarm optimization process, the fitness function is:

[0023]

[0024] Where F is the fitness function value, set L represents the vector set pointing from surface atoms to adsorbate molecules, set B is the set of all bonding guide vectors on adsorbate molecules, set P is the set of all pseudo-bond vectors, and the pseudo-bond vector is the vector pointing to the molecular adsorption site in the surface atom. Set L H The vectors in represent all potential hydrogen bonds in L\P, set N is the set of normal vectors of surface atoms, set V is the combination of elements in sets B, P, and N, U() is the artificial potential function considering the preference of adsorption sites, O() is the posture control function considering structural similarity, and U H () is the hydrogen bond potential function considering the influence of the micro-environment, B() is the steric hindrance function, and r0 is the threshold value of steric hindrance.

[0025] As a preferred technical solution, the artificial potential function is:

[0026]

[0027] Among them, U() is the artificial potential function, r ij is the distance between two atoms, ∈ ij , σ ij are the potential well depth and equilibrium distance of the atomic pair, i and j represent the atomic numbers of the two atoms, k is a parameter, U ′ () is the derivative function of the artificial potential function U().

[0028] As a preferred technical solution, the posture control function is:

[0029]

[0030] in, is the attitude control function, represents the bonding orientation vector of the adsorbate molecule, Represents the adsorption vector of the surface atoms.

[0031] As a preferred technical solution, the hydrogen bond potential function is:

[0032]

[0033] Among them, U H () is the hydrogen bond potential function, r ij is the distance between two atoms, ∈ ij ,σ ij is the potential well depth and equilibrium distance of the atomic pair.

[0034] Another aspect of the present invention provides an electronic device, comprising: one or more processors and a memory, wherein the memory stores one or more programs, and the one or more programs include instructions for executing the aforementioned heuristic multi-tooth adsorption configuration generation method based on particle swarm optimization.

[0035] Another aspect of the present invention provides a computer-readable storage medium comprising one or more programs for execution by one or more processors of an electronic device, wherein the one or more programs include instructions for executing the aforementioned heuristic multi-tooth adsorption configuration generation method based on particle swarm optimization.

[0036] Compared with the prior art, the present invention has at least one of the following beneficial effects:

[0037] (1) Considering various adsorption modes: The present invention introduces a particle swarm optimization (PSO) algorithm, which can fully consider the various adsorption postures of adsorbates at different active sites and take various adsorption configurations into consideration, thereby enhancing the modeling capability of complex surfaces and various adsorption modes of adsorbates.

[0038] (2) High generation efficiency and more reasonable generated adsorption configuration: The present invention realizes multi-dentate adsorption of adsorbates by optimizing the matching between adsorbates and surface atoms, making the generated structure more in line with chemical rules and avoiding the inefficiency of manual modeling. In the iterative process, structural similarity, adsorption site preference, and micro-environmental influence are fully considered to further improve the rationality of the generated adsorption configuration.

[0039] (3) Adaptability to complex systems: The present invention can effectively process complex catalytic surfaces such as clusters, step surfaces, etc., and realize multi-dentate adsorption of adsorbate molecules, overcoming the limitation that existing algorithms can only process simple surfaces and single-site adsorption. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a flow chart of a heuristic multi-tooth adsorption configuration generation method based on particle swarm optimization in an embodiment;

[0041] Figure 2 The adsorbent CHO and the catalyst ZnCrO are given in the example. x Schematic diagram of the process of generating a multidentate adsorption configuration using the example method. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0043] Example 1

[0044] In response to the problems existing in the aforementioned prior art, this embodiment provides a heuristic multi-dentate adsorption configuration generation method based on particle swarm optimization, which can automatically generate reasonable adsorption configurations on complex surfaces, especially considering the diverse adsorption modes of adsorbates on complex catalyst surfaces. By optimizing the matching mode between adsorbates and surface sites, the algorithm effectively solves the multi-dentate adsorption problem of complex adsorbates and improves the efficiency and rationality of initial configuration generation. The introduction of this technology will further improve the efficiency of automated modeling in computational chemistry and provide a solid data foundation for theoretical research on complex reaction systems.

[0045] For details, see Figure 1 , the method comprises the following steps:

[0046] Step S1, input of adsorbate and catalyst: firstly, select the target adsorbate molecule and catalyst material, and use the topological structure of the adsorbate molecule and the atomic coordinate information of the catalyst as the initial input.

[0047] Step S2, processing the adsorbate molecule: by comparing with the complete molecule in the molecule library, find the potential parent molecule of the adsorbate molecule. Further compare the topological structures of the two to determine the relative positions of atoms in the adsorbate molecule, adsorption sites and their bonding guide vectors.

[0048] Specifically, step S2 may include the following sub-steps:

[0049] Step S201, express the adsorbate molecule as a topological structure, and compare the topological structure with the complete molecule in the molecular library to determine a reasonable parent molecule. A reasonable parent molecule must meet the following conditions: its topological structure must contain the topological structure of the adsorbate molecule, that is, the topological map of the adsorbate is a subgraph of the topological map of the parent molecule. After matching with the parent molecule (i.e., after topological matching), the number of missing bonds of the adsorbate molecule should be less than or equal to 3, and all missing bonds must face the same side to ensure the rationality of the geometric configuration and chemical stability during the adsorption process. Since one adsorbate molecule may correspond to multiple parent molecules, resulting in a variety of possible adsorption postures, this algorithm will take them into consideration during the optimization process.

[0050] Step S202, a detailed comparison is made between the adsorbate molecule and the parent molecule, the relative positions of the atoms in the adsorbate molecule are extracted from the parent molecule, and the vector pointing from the missing bond position to the missing atom is defined as the bonding guide vector of the adsorbate molecule. These guide vectors not only reflect the bonding characteristics of the adsorbate molecule in the parent molecule, but also provide important geometric reference information for the subsequent optimization of the adsorption configuration. In this process, the chemical environment of the parent molecule is used as an initial reference for the adsorption of the adsorbate molecule on the catalyst surface, thereby ensuring the rationality of the adsorption configuration.

[0051] Step S3, treating the catalyst: scanning the catalyst material, identifying the surface atoms, further calculating the adsorption vectors of the surface atoms, and using them as potential adsorption sites.

[0052] Specifically, step S3 may include:

[0053] Step S301, using existing tools, such as Catkit software, to automatically identify the surface atomic positions of the catalytic material. For complex catalytic surfaces, users can manually specify the serial numbers of surface atoms during input to ensure that the algorithm accurately identifies all potential adsorption sites.

[0054] Step S302, for each surface atom, count its neighboring atoms within the set cutoff radius, and obtain the relative position vectors of these neighboring atoms and their corresponding distances. Then, weight the relative position vectors of the neighboring atoms according to the distance to calculate the adsorption vector of the surface atom. Finally, normalize the obtained adsorption vector as the adsorption vector of the surface atom. This adsorption vector takes into account the geometric characteristics of the surrounding environment of the surface atom and reflects its potential adsorption site characteristics, providing an important basis for the optimization of the subsequent adsorption configuration.

[0055] Step S4, molecular rigid body model construction: the adsorbate molecule is regarded as a rigid body with six degrees of freedom, and the spatial configuration of the adsorbate molecule is uniquely determined by three translation parameters and three rotation parameters in a fixed reference system.

[0056] Specifically, step S4 may include:

[0057] The entire adsorbate molecule is regarded as a rigid body model with six degrees of freedom, and its spatial position is described by three translation parameters (x, y, z) and three rotation parameters (a, b, c). In a fixed reference frame, these six parameters can uniquely determine the spatial configuration of the adsorbate molecule through the transformation of the special Euclidean group SE(3). Based on this, a bijective mapping relationship is established between the parameter set {S} and the atomic coordinate set {C}. This mapping realizes the unique determination of the molecular configuration through the SE(3) transformation, providing a clear geometric description framework for the optimization of the adsorption configuration.

[0058] Step S5, PSO algorithm optimization: using the particle swarm optimization algorithm, find the optimal adsorption configuration of the adsorbate molecular fragment on the catalyst surface in the high-dimensional parameter space, and determine the corresponding parameter group.

[0059] Specifically, step S5 may include the following sub-steps:

[0060] Step S501, considering structural similarity: This algorithm uses the posture control function To adjust the spatial position of the adsorbate molecule to ensure that it has a high similarity with the structure of its parent molecule when adsorbed on the catalyst surface. The formula of the posture control function is: in represents the bonding orientation vector of the adsorbate molecule, Represents the adsorption vector of the surface atoms.

[0061] Step S502, consider the adsorption site preference: In this embodiment, the Lennard-Jones potential function is used to comprehensively consider the adsorption bonding distance and the adsorption site selection preference. The Lennard-Jones potential parameters ∈ and σ are used to control the potential well depth and ideal bonding distance between different atomic pairs, where the potential well depth can be understood as the attraction preference between atomic pairs, and the ideal bonding distance can be approximated as the adsorption bonding distance. For specific systems, the ∈ values ​​of different atomic pairs can be adjusted according to practical experience to supplement and improve the deficiencies of the default parameters of the LJ potential function. Based on this, this algorithm introduces an artificial potential function U(r), whose formula is:

[0062]

[0063] Among them, r ij is the distance between two atoms, ∈ ij , σ ij are the potential well depth and equilibrium distance of the atomic pair respectively, i and j represent the atomic numbers of the two atoms. After testing, the PSO search efficiency is best when k=1.347.

[0064] Step S503, consider the influence of the micro-environment: In order to avoid obvious steric effect during adsorption, the algorithm introduces the steric function B(r, r0) and uses the hydrogen bond potential function U H (r) to stabilize the adsorption configuration of the adsorbate molecules. The expression of the steric hindrance function is: Where r represents the distance between two atoms, and r0 is a custom constant representing the threshold of steric hindrance, which is generally Hydrogen bond potential function U H The formula for (r) is:

[0065]

[0066] Among them, r ij is the distance between two atoms, ∈ ij ,σ ij is the potential well depth and equilibrium distance of the atomic pair, where i and j represent the atomic numbers.

[0067] Step S504, constructing a fitness function: comprehensively considering structural similarity, adsorption site preference and micro-environmental influence, the fitness function F is defined as:

[0068]

[0069] The set L represents the set of vectors pointing from the surface atoms to the atoms of the adsorbate molecules, and the set B is the set of all bonding guide vectors on the adsorbate molecules. j is the set of all bonding guide vectors on site j. During the optimization process, the vectors pointing to the molecular adsorption sites in the surface atoms are quasi-bond vectors. The set P is the set of all quasi-bond vectors. P must satisfy the following conditions at the same time: (1) The number of quasi-bond vectors is equal to the number of bonding guide vectors, that is, |P| = |B|, (2) for any and due in

[0070] Collection L H The vectors in represent all potential hydrogen bonds in L\P, mainly considering OH and NH bonds. The set N is the normal vector set of surface atoms, expressed as:

[0071] The set V is the combination of elements in the sets B, P, and N, and is defined as: In order to ensure that the bonding angle between the adsorbate molecules and the surface atoms is optimized, the set V must also make the combination The sum of cosines between them is minimum.

[0072] Step S6, adsorption structure restoration: according to the optimal parameter group obtained by the PSO algorithm, the optimal adsorption configuration of the adsorbate on the catalyst surface is reconstructed to complete the initialization process of the adsorption configuration.

[0073] In the following, we will use the adsorbent CHO and the catalyst ZnCrO x Take as an example the detailed process of generating multi-tooth adsorption configuration using this method is described. Figure 2, which mainly includes three steps: (1) Processing adsorbate molecules: first search for potential parent molecules of the adsorbate, then compare with the parent molecule to obtain the relative position of the atoms in the molecule and the adsorption guide vector; (2) Processing catalysts: identify the surface atoms of the catalyst and calculate the adsorption vectors of these surface atoms; (3) PSO optimization and adsorption configuration generation: through the PSO optimization algorithm, ensure that the adsorption configuration of the adsorbate is consistent with the parent molecule, and consider the preference and steric effect of the adsorbate molecule at the surface site, and finally generate a reasonable adsorption configuration. The picture shows two final adsorption configurations generated based on different parent molecules CH3OH and CH2O. The following will expand on the detailed description.

[0074] Step 1: Select CHO molecule as adsorbate and ZnCrO as catalyst x , input the topological structure information of CHO molecule and ZnCrO x Atomic coordinate information.

[0075] Step 2: Compare the topological structure of the CHO molecule with the complete molecules in the molecular library to identify its potential parent molecules, such as CH3OH, CH2O, etc. (see Figure 2 ). After identifying potential parent molecules, the topological structures of CHO and these parent molecules are further compared to obtain the relative positions of C, H, and O atoms in the CHO molecule, and accurately locate the missing bond sites in the CHO molecule. Through this process, the reasonable configuration of the CHO molecule and its adsorption guidance vector can be effectively determined.

[0076] According to this method, the adsorption orientation vector of CHO molecules is determined as:

[0077] Configuration 1 (with CH3OH as the parent molecule);

[0078] O-[0.28497,0.70564,-0.64874],

[0079] C-[-0.33548,-0.17959,-0.92477],

[0080] C-[-0.32349,0.89549,0.30569].

[0081] Configuration 2 (with CH2O as the parent molecule):

[0082] C-[-0.53409,-0.21559,0.81748].

[0083] It is worth noting that different parent molecules correspond to a variety of possible adsorption configurations of CHO molecules. Therefore, this method can effectively capture the various adsorption postures of CHO on the catalytic surface, thus providing a more complete theoretical basis for in-depth research on the interaction between adsorbates and catalyst surfaces.

[0084] Step 3: First identify ZnCrO x The surface atoms of the material, set a cutoff radius for each surface atom. Neighborhood range. Within this radius, all surrounding atoms within this range will be identified as neighboring atoms. For each surface atom, the relative position vector and the corresponding distance between it and all neighboring atoms are calculated. Next, the vector of each neighboring atom is weighted according to the distance to the surface atom. The farther the distance, the smaller the weight. By accumulating the weighted neighboring atom vectors, the adsorption vector of the surface atom can be obtained. Finally, the adsorption vector is normalized to standardize its length as the final adsorption vector of the surface atom.

[0085] The surface adsorption vector calculated by this method is as follows:

[0086] 48-O-[-0.00114,-0.00107,0.99999]

[0087] 49-O-[0.01705,-0.01142,0.99978]

[0088] 50-O-[0.00022,0.00038,0.99999]

[0089] 51-O-[0.01341,0.01662,0.99977]

[0090] 52-O-[-0.01099,-0.00910,0.99989]

[0091] 53-O-[0.00451,0.00661,0.99996]

[0092] 54-O-[-0.01989,-0.00605,0.99978]

[0093] 55-O-[-0.00979,0.00284,0.99994]

[0094] 56-O-[-0.02500,-0.00232,0.99968]

[0095] 57-O-[0.02208,0.00508,0.99974]

[0096] 58-O-[0.01228,0.00276,0.99992]

[0097] 59-O-[-0.01299,0.00460,0.99990]

[0098] 84-Cr-[-0.00321,0.00134,0.99999]

[0099] 85-Cr-[0.00847,0.00391,0.99995]

[0100] 86-Cr-[0.00551,-0.00165,0.99998]

[0101] 87-Cr-[-0.00807,-0.00623,0.99994]

[0102] 88-Cr-[-0.04822,0.00811,0.99880]

[0103] 89-Cr-[0.04820,-0.00520,0.99882]

[0104] 90-Cr-[0.01519,-0.00166,0.99988]

[0105] 91-Cr-[-0.02268,-0.00102,0.99974]

[0106] 104-Zn-[-0.05652,0.00305,0.99839]

[0107] 105-Zn-[0.04202,-0.00534,0.99910]

[0108] 106-Zn-[0.00862,-0.01059,0.99990]

[0109] 107-Zn-[-0.00517,0.00102,0.99998].

[0110] This method can effectively capture the structural information of the active center composed of surface atoms and their neighboring atoms, and quantify the adsorption preference of surface atoms.

[0111] Step 4: The relative positions of the atoms in the CHO molecule are obtained through step 2, so the CHO molecule can be regarded as a six-degree-of-freedom rigid body model, which consists of three translation parameters (x, y, z) and three rotation parameters (a, b, c). In a fixed reference system, these six degrees of freedom parameters can uniquely determine the spatial configuration of the molecule through the transformation of the special Euclidean group SE (3). Therefore, a vector containing six components [x1, x2, x3, x4, x5, x6] is initialized to describe the position and posture of the CHO molecule in three-dimensional space.

[0112] Step 5: Take the 6-dimensional vector describing the position of CHO molecules in step 4 as input and use the particle swarm optimization algorithm to optimize the parameters of the adsorption configuration. During the optimization process, the algorithm mainly considers the following aspects: (1) Structural similarity: CHO molecules in ZnCrO x The adsorption configuration of the surface should continuously approach the configuration of its parent molecule CH3OH, CH2O during the optimization process. (2) Preference for the selection of adsorption sites: The Lennard-Jones potential function is used to comprehensively consider the adsorption bonding distance and the preference for the selection of adsorption sites. In order to better adapt to the characteristics of a specific catalyst and adsorbate system, see Table 1. This embodiment adjusts the potential function parameters (such as ∈ value) of different atomic pairs based on practical experience to supplement and correct the default parameters of the LJ potential function to ensure that the model accurately describes the adsorption behavior. (3) Micro-environmental influence: In order to avoid obvious steric effects of the adsorption configuration, make full use of hydrogen bonding to stabilize the adsorption configuration of CHO.

[0113] Table 1 Relationship between potential function parameters and element 1 and element 2

[0114] Element 1 Element 2 ∈ value C Cr 8.0 C Zn 7.0 C O 4.0 O Cr 5.0 O Zn 6.0 O O 3.0 H Cr 3.0 H Zn 2.8 H O 2.0

[0115] Step 6: Figure 2 As shown in the figure, the parameter group obtained by the PSO algorithm is used to restore CHO in ZnCrO x Optimal adsorption structure on the surface.

[0116] In summary, the heuristic multi-tooth adsorption configuration generation method of this embodiment has the following advantages:

[0117] (1) Adaptability to complex systems: The algorithm can effectively process complex catalytic surfaces such as clusters and step surfaces, achieving multi-dentate adsorption of adsorbate molecules, overcoming the limitation of existing algorithms that can only process simple surfaces and single-site adsorption.

[0118] (2) Consideration of various adsorption modes: By introducing the particle swarm optimization (PSO) algorithm, this method can fully consider the various adsorption postures of the adsorbate at different active sites and take into account various adsorption configurations, thereby enhancing the modeling ability of complex surfaces and various adsorption modes of adsorbates.

[0119] (3) Improve generation efficiency and rationality: The algorithm achieves multi-dentate adsorption of adsorbates by optimizing the matching between adsorbates and surface atoms, making the generated structure more consistent with chemical rules and avoiding the inefficiency of manual modeling.

[0120] Example 2

[0121] This embodiment provides an electronic device, including: one or more processors and a memory, wherein the memory stores one or more programs, and the one or more programs include instructions for executing the heuristic multi-tooth adsorption configuration generation method based on particle swarm optimization as described in Example 1.

[0122] Example 3

[0123] This embodiment provides a computer-readable storage medium, including one or more programs for execution by one or more processors of an electronic device, wherein the one or more programs include instructions for executing the heuristic multi-tooth adsorption configuration generation method based on particle swarm optimization as described in Example 1.

[0124] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A heuristic multi-tooth adsorption configuration generation method based on particle swarm optimization, characterized in that: The steps include: Obtain topological structure information of adsorbate molecules and atomic coordinate information of catalyst materials; For the adsorbate molecule, the parent molecule is matched based on the topological structure information, and the bond-missing position in the parent molecule is used as the bonding guide vector of the adsorbate molecule, wherein the topological graph of the adsorbate molecule is a subgraph of the topological graph of the parent molecule; For the catalyst material, calculating the adsorption vectors of the surface atoms based on the atomic coordinate information; Establish a one-to-one mapping relationship between the rigid body model parameter group of the adsorbate molecule and the atomic coordinates within the adsorbate molecule; Based on the bonding orientation vector and the adsorption vector, searching for a rigid body model parameter group of an optimal adsorption configuration of adsorbate molecules on the catalyst surface by particle swarm optimization under the premise of considering at least one of structural similarity, adsorption site preference, and micro-environmental influence; Based on the rigid body model parameter group of the optimal adsorption configuration, the optimal adsorption configuration of the catalyst surface is restored using the mapping relationship.

2. The heuristic multi-tooth adsorption configuration generation method based on particle swarm optimization according to claim 1, characterized in that: The process of matching the parent molecule based on the topological structure information and using the missing bond position in the parent molecule as the bonding guide vector of the adsorbate molecule comprises the following steps: Convert the adsorbate molecule into a topological structure and match one or more parent molecules. After topological matching, the number of missing bonds of the adsorbate molecule is less than or equal to 3, and all missing bonds face the same side. The relative positions of atoms in the adsorbate molecule are extracted from the parent molecule, and the missing bond position is used as the bonding guide vector of the adsorbate molecule, wherein the bonding guide vector includes the bonding characteristic information and geometric reference information of the adsorbate molecule in the parent molecule.

3. The heuristic multi-tooth adsorption configuration generation method based on particle swarm optimization according to claim 1 is characterized in that: The process of calculating the adsorption vector of the surface atoms based on the atomic coordinate information comprises the following steps: Identify the locations of surface atoms of catalyst materials; For each surface atom, neighboring atoms are searched within the preset stage radius, the relative position vector and distance between neighboring atoms are calculated, the relative position vector is weighted according to the distance to obtain the adsorption vector of the surface atom and normalized.

4. The heuristic multi-tooth adsorption configuration generation method based on particle swarm optimization according to claim 1 is characterized in that: The construction process of the mapping relationship includes the following steps: Modeling the adsorbate molecule as a six-degree-of-freedom rigid body model determined by a parameter set, wherein the parameter set includes a translation parameter and a rotation parameter; Through the special Euclidean group SE(3) transformation, a one-to-one mapping relationship between the rigid body model parameter group and the atomic coordinates of the adsorbate molecules is established, and the spatial configuration of the adsorbate molecules is uniquely determined by the parameter group.

5. The heuristic multi-tooth adsorption configuration generation method based on particle swarm optimization according to claim 1, characterized in that: In the particle swarm optimization process, the fitness function is: Where F is the fitness function value, set L represents the vector set pointing from surface atoms to adsorbate molecules, set B is the set of all bonding guide vectors on adsorbate molecules, set P is the set of all pseudo-bond vectors, and the pseudo-bond vector is the vector pointing to the molecular adsorption site in the surface atom. Set L H The vectors in represent all potential hydrogen bonds in L\P, set N is the set of normal vectors of surface atoms, set V is the combination of elements in sets B, P, and N, U() is the artificial potential function considering the preference of adsorption sites, O() is the posture control function considering structural similarity, and U H () is the hydrogen bond potential function considering the influence of the micro-environment, B() is the steric hindrance function, and r0 is the threshold value of steric hindrance.

6. A heuristic multi-tooth adsorption configuration generation method based on particle swarm optimization according to claim 5, characterized in that: The artificial potential function is: Among them, U() is the artificial potential function, r ij is the distance between two atoms, ∈ ij , σ ij are the potential well depth and equilibrium distance of the atomic pair, i and j represent the atomic numbers of the two atoms, k is a parameter, U ′ () is the derivative function of the artificial potential function U().

7. The heuristic multi-tooth adsorption configuration generation method based on particle swarm optimization according to claim 5, characterized in that: The posture control function is: in, is the attitude control function, represents the bonding orientation vector of the adsorbate molecule, Represents the adsorption vector of the surface atoms.

8. The heuristic multi-tooth adsorption configuration generation method based on particle swarm optimization according to claim 5, characterized in that: The hydrogen bond potential function is: Among them, U H () is the hydrogen bond potential function, r ij is the distance between two atoms, ∈ ij ,σ ij is the potential well depth and equilibrium distance of the atomic pair.

9. An electronic device, characterized in that: include: One or more processors and a memory, wherein the memory stores one or more programs, wherein the one or more programs include instructions for executing the heuristic multi-tooth adsorption configuration generation method based on particle swarm optimization as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that: It comprises one or more programs for execution by one or more processors of an electronic device, wherein the one or more programs comprise instructions for executing the heuristic multi-tooth adsorption configuration generation method based on particle swarm optimization as claimed in any one of claims 1 to 8.

Images