Potential function construction method of complex-phase zirconia ceramic

Abstract

The invention provides a potential function construction method of complex-phase zirconia ceramic. The method comprises the following steps: respectively constructing crystal models of different components and different structures; performing simulation calculation on the constructed crystal model by using a density functional theory to obtain an initial training data set; performing fitting training on the initial training data set based on a deep potential energy learning model to obtain a plurality of initial potential function models; performing sampling simulation on the crystal model under a preset simulation condition by utilizing the plurality of initial potential function models, and screening and marking the initial potential function models based on a simulation result; and performing analog computation on the marked model by using a density functional theory, adding data obtained after analog computation into an initial training data set, and skipping to execute fitting training and screening marking steps by taking a new training data set as the initial training data set until a potential function model conforming to preset precision is generated. The potential function model in the scheme can accurately capture the temperature rise phase change process of the zirconia ceramic doped with the yttrium element.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of molecular simulation computing technology in material science, and in particular to a method for constructing a potential function of a multiphase zirconia ceramic. Background Art

[0002] At present, it is a common method to stabilize zirconium oxide by adding yttrium ions and other second-phase oxides. This treatment can stabilize the structure of tetragonal zirconium oxide to room temperature. Due to the great potential of zirconium-yttrium oxide ceramic material system, many laboratories have conducted extensive research on its phase transformation toughening mechanism.

[0003] In related technologies, the research on the phase transition behavior of zirconium-yttrium-oxide ceramic systems mainly relies on experimental methods. Although traditional experimental methods can intuitively observe the changes in the macroscopic properties of materials, it is often difficult to accurately analyze the microscopic mechanism of phase transition, such as the forces between atoms and the migration of oxygen vacancies. In addition, due to the limitations of experimental equipment and conditions, many experiments are difficult to fully reproduce the phase transition process of materials under actual application conditions.

[0004] In addition, traditional atomic simulation methods mostly rely on classical potential function models, which have obvious limitations in capturing the phase change behavior of complex oxide ceramic materials and find it difficult to accurately describe complex interactions, especially when considering yttrium doping, whose effects on the material's structural stability and phase change behavior are often simplified or ignored.

[0005] Therefore, there is an urgent need to provide a potential function construction method for multiphase zirconia ceramics. Summary of the invention

[0006] In order to solve the problem that the traditional potential function model is difficult to accurately capture the temperature-increasing phase change process of zirconia ceramics when yttrium is doped into zirconia ceramics, an embodiment of the present invention provides a potential function construction method for multiphase zirconia ceramics.

[0007] In a first aspect, an embodiment of the present invention provides a method for constructing a potential function of a multiphase zirconia ceramic, the method comprising:

[0008] Step S1: constructing crystal models of different components and different structures respectively; wherein the crystal models include a single-component crystal model, a two-component crystal model and a three-component crystal model respectively formed by zirconium, oxygen and yttrium;

[0009] Step S2: using density theory to simulate and calculate the constructed crystal model to obtain an initial training data set;

[0010] Step S3: performing fitting training on the initial training data set based on the deep potential energy learning model to obtain multiple initial potential function models;

[0011] Step S4: using a plurality of initial potential function models to perform sampling simulation on the crystal model under preset simulation conditions, and screening and marking the initial potential function models based on the simulation results;

[0012] Step S5: Use density functional theory to simulate the marked model, add the data obtained after the simulation calculation to the initial training data set to form a new training data set, and use the new training data set as the initial training data set to jump to step S3 to step S5 until a potential function model that meets the preset accuracy is generated.

[0013] In a second aspect, an embodiment of the present invention further provides a device for constructing a potential function of a multiphase zirconia ceramic, the device comprising:

[0014] A crystal construction unit, used to construct crystal models of different components and different structures respectively; wherein the crystal models include a single-component crystal model, a two-component crystal model and a three-component crystal model respectively formed by zirconium, oxygen and yttrium;

[0015] A simulation calculation unit, used for performing simulation calculation on the constructed crystal model using density functional theory to obtain an initial training data set;

[0016] A fitting training unit, used for performing fitting training on the initial training data set based on a deep potential energy learning model to obtain multiple initial potential function models;

[0017] A screening and marking unit, used to perform sampling simulation on the crystal model under preset simulation conditions using a plurality of initial potential function models, and to screen and mark the initial potential function models based on the simulation results;

[0018] The loop unit is used to use density functional theory to simulate the marked model, add the data obtained after the simulation calculation to the initial training data set to form a new training data set, and use the new training data set as the initial training data set to jump to the fitting training and screening marking steps until a potential function model that meets the preset accuracy is generated.

[0019] In a third aspect, an embodiment of the present invention further provides a computing device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method described in any embodiment of this specification is implemented.

[0020] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, enables the computer to execute the method described in any embodiment of this specification.

[0021] On the other hand, an embodiment of the present invention further provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the method for constructing a potential function of a multiphase zirconia ceramic as described above.

[0022] The embodiment of the present invention provides a method for constructing a potential function of a multiphase zirconia ceramic. First, a single-component crystal model, a two-component crystal model and a three-component crystal model respectively formed by zirconium, oxygen and yttrium are constructed, and then each constructed crystal model is simulated and calculated using density functional theory, so as to obtain the corresponding relationship between different crystal models and energy, atomic force and displacement, thereby obtaining a database that can characterize zirconia ceramics under different yttrium doping concentrations, and the database is used as an initial training data set for fitting training using a deep potential energy learning model, and the initial training data set is expanded in combination with the results obtained by density functional theory calculation during the training process for iterative training, thereby improving the adaptability and accuracy of the potential function model; in this way, the generated potential function model can accurately capture the temperature-increasing phase change process of yttrium-doped zirconia ceramics, while making up for the disadvantage of low efficiency of first-principles calculations, and achieving calculation efficiency close to that of empirical potential functions. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 creative work.

[0024] Figure 1 It is a flow chart of a method for constructing a potential function of a multiphase zirconia ceramic provided by one embodiment of the present invention;

[0025] Figure 2 It is a fitting training flow chart of a potential function of a multiphase zirconia ceramic provided by an embodiment of the present invention;

[0026] Figure 3 This is a comparison diagram of the potential function model of the yttrium-doped zirconia ceramic provided by the embodiment of the present invention and the DFT calculation result;

[0027] Figure 4 is a hardware architecture diagram of a computing device provided by an embodiment of the present invention;

[0028] Figure 5 It is a structural diagram of a potential function construction device for a multiphase zirconia ceramic provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0030] As mentioned above, in view of the experimental limitations and defects of traditional atomic simulation methods in the study of the phase transition behavior of yttrium-doped zirconia ceramics, the present invention considers constructing a potential function suitable for the zirconium-yttrium oxide ceramic system based on a machine learning potential field model to accurately simulate and predict its phase transition behavior.

[0031] Please refer to Figure 1 The embodiment of the present invention provides a method for constructing a potential function of a multiphase zirconia ceramic, the method comprising:

[0032] Step S1: constructing crystal models of different components and different structures respectively; wherein the crystal models include a single-component crystal model, a two-component crystal model and a three-component crystal model respectively formed by zirconium, oxygen and yttrium;

[0033] Step S2: using density functional theory to perform simulation calculations on the constructed crystal model to obtain an initial training data set;

[0034] Step S3: performing fitting training on the initial training data set based on the deep potential energy learning model to obtain multiple initial potential function models;

[0035] Step S4: using a plurality of initial potential function models to perform sampling simulation on the crystal model under preset simulation conditions, and screening and marking the initial potential function models based on the simulation results;

[0036] Step S5: Use density functional theory to simulate the marked model, add the data obtained after the simulation calculation to the initial training data set to form a new training data set, and use the new training data set as the initial training data set to jump to step S3 to step S5 until a potential function model that meets the preset accuracy is generated.

[0037] In an embodiment of the present invention, a single-component crystal model, a two-component crystal model and a three-component crystal model respectively formed by zirconium, oxygen and yttrium are first constructed, and then density functional theory is used to simulate and calculate each constructed crystal model, so as to obtain the corresponding relationship between different crystal models and energy, atomic force and displacement, thereby obtaining a database that can characterize zirconia ceramics under different yttrium doping concentrations, and the database is used as an initial training data set for fitting training using a deep potential energy learning model, and the initial training data set is expanded in combination with the results obtained by density functional theory calculation during the training process for iterative training, thereby improving the adaptability and accuracy of the potential function model; in this way, the generated potential function model can accurately capture the temperature-raising phase change process of yttrium-doped zirconia ceramics, while making up for the disadvantage of low efficiency of first-principles calculations, and achieving calculation efficiency close to that of the empirical potential function.

[0038] For step S1:

[0039] In some embodiments, the single-component crystal model includes a zirconium element crystal model and an oxygen element crystal model, the two-component crystal model includes a zirconium oxide crystal model and a yttrium oxide crystal model, and the three-component crystal model includes a zirconium oxide doped with yttrium element crystal model; the crystal model includes a transition state structure, a surface structure and a defect structure.

[0040] In some embodiments, the crystal model is constructed by:

[0041] Based on the first principles software, different types of zirconium oxide crystal models, yttrium oxide crystal models, zirconium element crystal models and oxygen element crystal models are respectively constructed to obtain a single-component crystal model and a two-component crystal model; wherein the zirconium oxide crystal model includes crystal models of Pbca-ZrO2, Pmna-ZrO2, m-ZrO2, t-ZrO2, c-ZrO2 and Pbc21-ZrO2;

[0042] Based on the zirconium oxide crystal model, a three-component crystal model based on at least two types of zirconium oxide crystal models doped with different concentrations of yttrium atoms is constructed;

[0043] According to the single-component crystal model, the two-component crystal model and the three-component crystal model, the transition state structure, the surface structure and the defect structure of each crystal model are constructed respectively.

[0044] In an embodiment of the present invention, different types of zirconium oxide crystal models are first constructed by first-principles software, and then based on two types of the crystal models (such as t-ZrO2 and c-ZrO2 crystal models), different concentrations of yttrium atoms (for example, 2.84mol%, 5.68mol%, 8.53mol%, 11.36mol%) are doped into the t-ZrO2 and c-ZrO2 crystal models, respectively, to obtain a plurality of zirconium oxide crystal models doped with yttrium elements (three-component crystal models), and further on the basis of the above crystal models, crystal models of the transition state structure, surface structure and defect structure of each crystal model are constructed, and a relatively comprehensive initial training data set can be formed based on the different crystal models constructed above, and then based on the data set, it is beneficial to train a potential function model that can accurately reflect the phase change process of zirconium oxide doped with yttrium elements.

[0045] For step S2:

[0046] In some embodiments, step S2 includes: using density functional theory to perform structural optimization, perturbation and ab initio molecular dynamics simulation on each constructed crystal model in turn, to obtain the energy, atomic force and displacement of each crystal model, so as to further obtain an initial training data set.

[0047] In this embodiment, density generalization theory (DFT) is used to simulate and calculate each constructed crystal model to obtain the corresponding relationship between the above-mentioned different crystal models and the energy, atomic force and displacement, and the DeePMD-kit software package is used to extract the energy, atomic force and displacement information of all crystal models in the result file to construct an initial data set for training the potential function, which is conducive to obtaining a potential function model that can accurately reflect the phase change process of different crystal models.

[0048] For steps S3 to S5:

[0049] In some embodiments, before performing fitting training, the method further includes converting the initial training set into .raw and .npy formats using the DeePMD kit software package.

[0050] In some implementations, during the fitting training process, the control file is in .json format, and the hyperparameters are set as follows:

[0051] Set the descriptor type with attention mechanism to se_atten_V2; set the rcut parameter to 9.0 and the rut_smth parameter to 8.5; set the size of the embedded network neuron to [25, 50, 100] and the size of the fitted network to [240, 240, 240].

[0052] In the embodiment of the present invention, before performing fitting training, the initial training set is first converted into .raw and .npy formats using the DeePMD-kit software package, such as Figure 2 As shown, the DPGEN potential function network based on the TensorFlow framework is then used to fit the initial training data set through an active learning strategy to construct a comprehensive crystal model - a potential function model of energy, atomic force and displacement. During the training process, the rcut parameter is set to 9.0 and the rut_smth parameter is set to 8.5 to reasonably affect the atomic interactions within the cutoff radius.

[0053] In the active learning mode of the DPGEN potential function network, multiple machine learning potential function models are used as the potential for atomic interaction. Under preset simulation conditions (temperature and pressure), the initial potential function model is used to sample and simulate multiple crystal structure models constructed in step S1, and the maximum deviation value of the atomic force of each crystal model is obtained, and the predicted deviation range of the machine learning potential function model is set as the upper and lower limits of the error criterion. According to the error criterion, the model is screened within this range, and the model exceeding the error criterion is eliminated. The model within the error criterion range is regarded as a model with hope for improvement and marked, and the marked initial potential function model is re-calculated (density functional theory) DFT, and then the calculated results are added to the initial training data set, and steps S2 and S5 are repeated until the iteration converges, so as to obtain a potential function model that meets the preset accuracy. Finally, by adjusting the weights of energy, atomic force and displacement, the potential function model that meets the preset accuracy is trained for a long number of steps to obtain the final potential function model.

[0054] In the embodiment of the present invention, the potential function model (.pb file) generated in a loop is compared with the validation set data (calculated by density general theory (DFT) simulation of the constructed crystal model) by using the dptest command in the DeePMD-kit software package (comparing physical quantities such as energy, atomic force and displacement under DFT accuracy) to evaluate the accuracy of the potential function model. Furthermore, the potential function model is used in the LAMMPS molecular dynamics software to perform a series of simulations, including equation of state, radial distribution function, crystal axis length, and temperature rise and fall relaxation, and the calculation results are compared with the results at the DFT level, such as Figure 3 It can be seen from the figure that the potential function model in the embodiment of the present invention has good stability and accuracy.

[0055] In summary, in the embodiments of the present invention, by comparing the DFT calculation results with the prediction results of the machine learning potential function model, an error criterion is established to optimize the model, thereby ensuring the stability and accuracy of the model, enabling the potential function to accurately capture the complex atomic interactions in the zirconium yttrium oxide ceramic system, and solving the problem of insufficient simulation accuracy of yttrium-doped zirconium oxide ceramics in the phase change process. At the same time, it makes up for the shortcoming of low efficiency of first-principles calculations and achieves calculation efficiency close to that of the empirical potential function.

[0056] like Figure 4 , Figure 5 As shown, an embodiment of the present invention provides a potential function construction device for a multiphase zirconia ceramic. The device embodiment can be implemented by software, or by hardware or a combination of software and hardware. From the hardware level, Figure 4 As shown, it is a hardware architecture diagram of a computing device where a potential function construction device of a multiphase zirconia ceramic provided by an embodiment of the present invention is located, except Figure 4 In addition to the processor, memory, network interface, and non-volatile memory shown in the figure, the computing device in which the device is located in the embodiment may also generally include other hardware, such as a forwarding chip responsible for processing messages, etc. Taking software implementation as an example, Figure 5 As shown, as a device in a logical sense, the CPU of the computing device in which it is located reads the corresponding computer program in the non-volatile memory into the memory and runs it. This embodiment provides a potential function construction device for a multiphase zirconia ceramic, the device comprising:

[0057] The crystal construction unit 501 is used to construct crystal models of different components and different structures respectively; wherein the crystal models include a single-component crystal model, a two-component crystal model and a three-component crystal model respectively formed by zirconium, oxygen and yttrium;

[0058] A simulation calculation unit 502 is used to perform simulation calculation on the constructed crystal model using density functional theory to obtain an initial training data set;

[0059] A fitting training unit 503 is used to perform fitting training on the initial training data set based on a deep potential energy learning model to obtain multiple initial potential function models;

[0060] A screening and marking unit 504 is used to perform sampling simulation on the crystal model under preset simulation conditions using multiple initial potential function models, and to screen and mark the initial potential function models based on the simulation results;

[0061] The loop unit 505 is used to use density functional theory to simulate the marked model, add the data obtained after the simulation calculation to the initial training data set to form a new training data set, and use the new training data set as the initial training data set to jump to the fitting training and screening marking steps until a potential function model that meets the preset accuracy is generated.

[0062] In one embodiment of the present invention, in the crystal building unit 501, the single-component crystal model includes a zirconium element crystal model and an oxygen element crystal model, the two-component crystal model includes a zirconium oxide crystal model and an yttrium oxide crystal model, and the three-component crystal model includes a zirconium oxide doped yttrium element crystal model; the crystal model includes a transition state structure, a surface structure and a defect structure.

[0063] In one embodiment of the present invention, in the crystal construction unit 501, the crystal model is constructed in the following manner:

[0064] Based on the first principles software, different types of zirconium oxide crystal models, yttrium oxide crystal models, zirconium element crystal models and oxygen element crystal models are respectively constructed to obtain a single-component crystal model and a two-component crystal model; wherein the zirconium oxide crystal model includes crystal models of Pbca-ZrO2, Pmna-ZrO2, m-ZrO2, t-ZrO2, c-ZrO2 and Pbc21-ZrO2;

[0065] Based on the zirconium oxide crystal model, a three-component crystal model based on at least two types of zirconium oxide crystal models doped with different concentrations of yttrium atoms is constructed;

[0066] According to the single-component crystal model, the two-component crystal model and the three-component crystal model, the transition state structure, the surface structure and the defect structure of each crystal model are constructed respectively.

[0067] In one embodiment of the present invention, when the simulation calculation unit 502 uses density functional theory to perform simulation calculation on the constructed crystal model to obtain an initial training data set, it includes:

[0068] Density functional theory is used to perform structural optimization, perturbation and ab initio molecular dynamics simulation on each constructed crystal model in turn to obtain the energy, atomic force and displacement of each crystal model, so as to further obtain the initial training data set.

[0069] In one embodiment of the present invention, in the fitting training unit 503, before performing the fitting training, the step of converting the initial training set into .raw and .npy formats using the DeePMD-kit software package is also included.

[0070] In one embodiment of the present invention, in the fitting training unit 503, during the fitting training process, the control file adopts the .json format, and the hyper parameters are set as follows:

[0071] Set the descriptor type with attention mechanism to se_atten_V2; set the rcut parameter to 9.0 and the rut_smth parameter to 8.5; set the size of the embedded network neuron to [25, 50, 100] and the size of the fitted network to [240, 240, 240].

[0072] It is understood that the structure illustrated in the embodiment of the present invention does not constitute a specific limitation on a potential function construction device for a multiphase zirconia ceramic. In other embodiments of the present invention, a potential function construction device for a multiphase zirconia ceramic may include more or fewer components than shown in the figure, or combine certain components, or split certain components, or arrange the components differently. The components shown in the figure may be implemented in hardware, software, or a combination of software and hardware.

[0073] The information interaction, execution process and other contents between the modules in the above-mentioned device are based on the same concept as the embodiment of the method of the present invention. For the specific contents, please refer to the description in the embodiment of the method of the present invention, and no further description is given here.

[0074] An embodiment of the present invention further provides a computing device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, a potential function construction method for a multiphase zirconia ceramic in any embodiment of the present invention is implemented.

[0075] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the processor executes a potential function construction method for a multiphase zirconia ceramic in any embodiment of the present invention.

[0076] Specifically, a system or device equipped with a storage medium can be provided, on which software program code that implements the functions of any of the above-mentioned embodiments is stored, and a computer (or CPU or MPU) of the system or device can be enabled to read and execute the program code stored in the storage medium.

[0077] In this case, the program code itself read from the storage medium can realize the function of any one of the above-mentioned embodiments, and thus the program code and the storage medium storing the program code constitute a part of the present invention.

[0078] The storage medium embodiments for providing the program code include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), a magnetic tape, a non-volatile memory card, and a ROM. Alternatively, the program code can be downloaded from a server computer by a communication network.

[0079] In addition, it should be clear that the functions of any of the above embodiments can be implemented not only by executing the program code read by the computer, but also by enabling an operating system operating on the computer to complete part or all of the actual operations based on instructions from the program code.

[0080] In addition, it can be understood that the program code read from the storage medium is written to a memory provided in an expansion board inserted into the computer or to a memory provided in an expansion module connected to the computer, and then based on the instructions of the program code, a CPU installed on the expansion board or expansion module is enabled to perform part or all of the actual operations, thereby realizing the functions of any of the above-mentioned embodiments.

[0081] An embodiment of the present application also provides a computer-readable storage medium, on which is stored at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by a processor to implement a potential function construction method for a multiphase zirconia ceramic provided in the above-mentioned method embodiments.

[0082] It should be noted that, in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0083] A person of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiments; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk, etc., various media that can store program codes.

[0084] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for constructing a potential function of a multiphase zirconia ceramic, characterized in that: include: Step S1: constructing crystal models of different components and different structures respectively; wherein the crystal models include a single-component crystal model, a two-component crystal model and a three-component crystal model respectively formed by zirconium, oxygen and yttrium; Step S2: using density functional theory to perform simulation calculations on the constructed crystal model to obtain an initial training data set; Step S3: performing fitting training on the initial training data set based on the deep potential energy learning model to obtain multiple initial potential function models; Step S4: using a plurality of initial potential function models to perform sampling simulation on the crystal model under preset simulation conditions, and screening and marking the initial potential function models based on the simulation results; Step S5: Use density functional theory to simulate the marked model, add the data obtained after the simulation calculation to the initial training data set to form a new training data set, and use the new training data set as the initial training data set to jump to step S3 to step S5 until a potential function model that meets the preset accuracy is generated.

2. The method according to claim 1, characterized in that The single-component crystal model includes a zirconium element crystal model and an oxygen element crystal model, the two-component crystal model includes a zirconium oxide crystal model and a yttrium oxide crystal model, and the three-component crystal model includes a zirconium oxide doped with yttrium element crystal model; the crystal model includes a transition state structure, a surface structure and a defect structure.

3. The method according to claim 2, characterized in that The crystal model is constructed in the following way: Based on the first principles software, different types of zirconium oxide crystal models, yttrium oxide crystal models, zirconium element crystal models and oxygen element crystal models are respectively constructed to obtain a single-component crystal model and a two-component crystal model; wherein the zirconium oxide crystal model includes crystal models of Pbca-ZrO2, Pmna-ZrO2, m-ZrO2, t-ZrO2, c-ZrO2 and Pbc21-ZrO2; Based on the zirconium oxide crystal model, a three-component crystal model based on at least two types of zirconium oxide crystal models doped with different concentrations of yttrium atoms is constructed; According to the single-component crystal model, the two-component crystal model and the three-component crystal model, the transition state structure, the surface structure and the defect structure of each crystal model are constructed respectively.

4. The method according to claim 1, characterized in that: The method uses density functional theory to simulate and calculate the constructed crystal model to obtain an initial training data set, including: Density functional theory is used to perform structural optimization, perturbation and ab initio molecular dynamics simulation on each constructed crystal model in turn to obtain the energy, atomic force and displacement of each crystal model, so as to further obtain the initial training data set.

5. The method according to claim 1, characterized in that Before performing fitting training, the method further includes converting the initial training set into .raw and .npy formats using the DeePMD-kit software package.

6. The method according to any one of claims 1 to 5, characterized in that During the fitting training process, the control file uses the .json format and sets the hyperparameters as follows: Set the descriptor type with attention mechanism to se_atten_V2; set the rcut parameter to 9.0 and the rut_smth parameter to 8.5; set the size of the embedded network neuron to [25, 50, 100] and the size of the fitted network to [240, 240, 240].

7. A potential function construction device for multiphase zirconia ceramics, characterized in that: The device comprises: A crystal construction unit, used to construct crystal models of different components and different structures respectively; wherein the crystal models include a single-component crystal model, a two-component crystal model and a three-component crystal model respectively formed by zirconium, oxygen and yttrium; A simulation calculation unit, used for performing simulation calculation on the constructed crystal model using density functional theory to obtain an initial training data set; A fitting training unit, used for performing fitting training on the initial training data set based on a deep potential energy learning model to obtain multiple initial potential function models; A screening and marking unit, used to perform sampling simulation on the crystal model under preset simulation conditions using a plurality of initial potential function models, and to screen and mark the initial potential function models based on the simulation results; The loop unit is used to use density functional theory to simulate the marked model, add the data obtained after the simulation calculation to the initial training data set to form a new training data set, and use the new training data set as the initial training data set to jump to the fitting training and screening marking steps until a potential function model that meets the preset accuracy is generated.

8. A computing device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to execute the method according to any one of claims 1 to 6.

10. A computer program product, characterized in that The method comprises a computer program, wherein when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images