Design method of multi-principal element ceramic material for laser protection

Abstract

The invention discloses a design method of a multi-principal-element ceramic material for laser protection, and belongs to the technical field of ceramic materials, an ATAT tool is adopted to construct an atomic structure of the multi-principal-element ceramic material, and the atomic structure comprises two sets of FCC structure nests composed of 40 atoms. A data set containing various configurations such as stretching, compression, perturbation and the like is generated through calculation of a first principle. A deep potential (DP) model and a DPGEN training tool are used, and in a plurality of iteration cycles, a machine learning potential function is gradually optimized by utilizing three stages of exploration, marking and training. Through the iterative training process, an accurate potential function model can be obtained and is used for efficiently predicting physical properties such as energy, force and pressure of a material. According to the method, the design efficiency of the multi-principal-element ceramic material is improved, the performance of the multi-principal-element ceramic material in laser radiation, high temperature and other extreme environments can be optimized, and the method particularly has wide application prospects in the field of laser protection.

Description

Technical Field

[0001] The invention relates to the technical field of ceramic materials, and in particular to a design method of a multi-principal-element ceramic material for laser protection. Background Art

[0002] With the rapid development of high-power laser technology, laser weapons, laser detection, laser communications and other fields are increasingly widely used. In order to cope with the destructive effects of laser radiation on materials, especially in the fields of military, aerospace and high-energy physics experiments, the research and development of laser protection materials has become particularly important. Laser protection materials need to have excellent heat resistance, radiation resistance, mechanical strength and chemical stability under high temperature, strong radiation and severe thermal stress.

[0003] In recent years, high entropy ceramics (HECs), as a new type of material system, have attracted widespread attention due to their excellent performance in extreme environments such as high temperature, corrosion resistance, and radiation resistance. High entropy ceramics are composed of multiple main elements and can maintain good structural stability in a wider temperature range. At the same time, their complex composition and structure enable them to exhibit stronger thermal stability and damage resistance when facing extreme conditions such as laser radiation.

[0004] However, despite the great potential of high-entropy ceramic materials, their design and optimization process still faces many challenges. Traditional ceramic material design methods often rely on empirical methods or experimental-based step-by-step optimization, which not only consumes a lot of time and resources, but also makes it difficult to efficiently predict and optimize the performance of complex multi-principal material systems. Therefore, how to design high-entropy ceramic materials with excellent laser protection performance has become an important problem to be solved in the current field of materials science. In order to overcome the above challenges, computational material design methods have gradually become an emerging direction in materials research. Through first-principles calculations (such as calculations based on density functional theory DFT) and molecular dynamics simulations (MD), the physical properties and behaviors of materials can be accurately predicted at the atomic scale, which provides a powerful tool for the design and performance optimization of new materials. However, although these methods can provide very accurate simulation results, their computational cost is high, and when faced with large-scale multi-component, multi-phase complex materials, there is still a problem of low computational efficiency.

[0005] In recent years, with the development of machine learning technology, machine learning models, especially machine learning potentials (MLPs), have been proposed to replace traditional force field models, significantly improving computational efficiency and accuracy. By learning a large amount of data obtained from first-principles calculations, machine learning potentials can accurately predict physical quantities such as energy, force, and pressure of materials, and can be used for large-scale material simulation and optimization. This makes the application of machine learning methods in multi-component systems and complex material design a very promising direction. Despite this, there are still some limitations in the development of existing machine learning potential functions, such as insufficient quality and quantity of data sampling, incompleteness of training sets, and immature efficient potential function optimization methods. Therefore, how to combine first-principles calculations with machine learning methods to develop efficient, accurate, and applicable machine learning potential functions for complex multi-principal ceramic materials is still an important research topic in the field of material design. Summary of the invention

[0006] The purpose of the present invention is to provide a design method for multi-principal component ceramic materials for laser protection, which combines molecular dynamics simulation, first-principles calculation and machine learning potential function to optimize the structure and performance of ceramic materials through efficient computing means.

[0007] This paper introduces a combination of molecular dynamics and machine learning potential functions to propose a new and efficient design method for the development of multi-principal ceramic materials for laser protection. By using the DPGEN tool to train the deep potential model, combined with the data set generation and iterative optimization of first-principles calculations, the physical properties of multi-principal ceramic materials under different extreme conditions can be accurately predicted, thus providing a new computational approach for the design and optimization of laser protection materials.

[0008] To achieve the above object, the present invention provides a design method for a multi-principal-element ceramic material for laser protection, the method comprising the following steps:

[0009] 1) The atomic structure of high entropy ceramics is constructed by using the ATAT tool. The structure includes 40 atoms, including titanium, zirconium, hafnium, gadolinium and tantalum elements, and two sets of face-centered cubic structures are nested to obtain a model of multi-principal ceramic materials;

[0010] 2) Generate stress-strain data, energy and force data of multi-principal ceramic materials based on first-principles calculations, and construct training data sets for machine learning, including tensile, compressive and perturbation configurations;

[0011] 3) Generate machine learning potential functions using the DPGEN tool, in which atomic positions are mapped to energy and forces through a deep potential model, and the potential function is optimized through an iterative training process;

[0012] 4) Use the optimized machine learning potential function to predict and optimize the thermodynamic and mechanical properties of multi-principal component ceramic materials, including the thermal expansion coefficient, Young's modulus, shear modulus, bulk modulus, yield strength, fracture toughness and laser radiation protection capability of the material.

[0013] Preferably, the multi-principal-component ceramic material obtained after calculation and optimization using the above method is (Ti0.2Zr0.2Hf0.2Gd0.2Ta0.2)C, which has a high yield strength and excellent plastic deformation ability, and can maintain a stable structure under high-energy laser irradiation.

[0014] Preferably, the machine learning potential function is optimized through a repeated iterative training process, which includes an exploration phase, a labeling phase and a training phase to improve the quality of the training set and the prediction accuracy of the potential function.

[0015] Preferably, the prediction of thermodynamic and mechanical properties of multi-principal component ceramic materials also includes the mechanical response of the materials under laser radiation, high temperature, compression and tension conditions.

[0016] Preferably, the multi-principal component ceramic material has a Young's modulus of 349.53 GPa, a shear modulus of 141.94 GPa, a bulk modulus of 216.79 GPa, a Poisson's ratio of 0.23, and a maximum stress of 14.4987 GPa.

[0017] Preferably, through the machine learning potential function, the thermal expansion coefficient of the (Ti0.2Zr0.2Hf0.2Gd0.2Ta0.2)C material at a temperature of 400K to 1000K can be predicted, and the material has stable thermal expansion behavior at high temperatures.

[0018] Preferably, the multi-principal-component ceramic material can effectively cope with high-power laser irradiation and high-temperature environments, and is suitable for laser weapons, protective equipment, aerospace, and nuclear energy fields.

[0019] Therefore, the present invention adopts a design method of a multi-principal ceramic material for laser protection of the above structure, including steps from modeling of the atomic structure of the material to machine learning model training, material performance prediction, etc., to efficiently optimize the thermodynamic and mechanical properties of the (Ti0.2Zr0.2Hf0.2Gd0.2Ta0.2)C ceramic material, and ensure its application under extreme conditions such as laser protection. The obtained (Ti0.2Zr0.2Hf0.2Gd0.2Ta0.2)C high entropy ceramic material can maintain good mechanical stability and damage resistance under extreme environments such as laser radiation and high temperature. The higher bulk modulus and shear modulus ensure that the material can effectively resist pressure under laser radiation, while the higher Young's modulus ensures that the material will not undergo significant deformation under the action of high-power lasers. In addition, the moderate Poisson's ratio further improves the mechanical stability of the material.

[0020] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A flow chart of a design method of a multi-principal-element ceramic material for laser protection according to the present invention;

[0022] Figure 2 It is a schematic diagram of the structure of the multi-principal-element ceramic material of the present invention;

[0023] Figure 3 The stress response of the multi-principal-element ceramic material of the present invention under different strain conditions;

[0024] Figure 4 The temperature dependence diagram of the thermal expansion coefficient of the multi-principal element ceramic material of the present invention;

[0025] Figure 5 This is a flow chart of the training of the potential function of the multi-principal component ceramic material machine learning of the present invention;

[0026] Figure 6 The energy and force loss function graphs of the potential function for machine learning of the present invention;

[0027] Figure 7 This is a comparison chart between the energy and force test of the machine learning potential function of the present invention and DFT. DETAILED DESCRIPTION

[0028] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods. Obviously, the described embodiments are only 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 making creative work are within the scope of protection of the present invention.

[0029] The terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms "a", "said" and "the" used in the embodiments of the present invention and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings, and "multiple" generally includes at least two.

[0030] It should also be noted that the term "includes", "comprising" or any other variation thereof is intended to cover non-exclusive inclusion, so that a commodity 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 commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprising a ..." do not exclude the existence of other identical elements in the commodity or device including the elements.

[0031] See attached Figure 1-7 As shown, the present application provides a design method for a multi-principal-element ceramic material for laser protection, the method comprising the following steps:

[0032] 1) The atomic structure of high entropy ceramics is constructed by using the ATAT tool. The structure includes 40 atoms, including titanium, zirconium, hafnium, gadolinium and tantalum elements, and two sets of face-centered cubic structures are nested to obtain a model of multi-principal ceramic materials;

[0033] Specifically, ATAT (Alloy Theoretic Automated Toolkit) is used to model the atomic structure of high entropy ceramic materials. The ATAT tool can simulate and generate the crystal structure of materials with multiple constituent elements. In order to ensure the design accuracy of multi-principal ceramic materials, 40 atoms were selected and two sets of face-centered cubic (FCC) structure nested models were constructed. This structural design helps the material maintain stability under high temperature and laser irradiation.

[0034] 2) Generate stress-strain data, energy and force data of multi-principal ceramic materials based on first-principles calculations, and construct training data sets for machine learning, including tensile, compressive and perturbation configurations;

[0035] Specifically, based on the constructed atomic model, first-principles calculations (DFT) were used to calculate the thermodynamic properties, mechanical properties, and force and energy responses of the material. To cover the behavior of the material under different stress conditions, a variety of configurations were generated, including:

[0036] Tensile configuration: simulates the performance of materials under tensile stress in different directions;

[0037] Compression configuration: study the response and mechanical properties of materials under external pressure;

[0038] Perturbation configuration: simulate small-amplitude perturbations between atoms to analyze the stability of materials and atomic mechanical behavior.

[0039] Through these first-principles calculations, we obtained physical quantities such as energy, force, and pressure of materials under different structures. These data provide rich training samples for subsequent machine learning potential training.

[0040] 3) Generate machine learning potential functions using the DPGEN tool, in which atomic positions are mapped to energy and forces through a deep potential model, and the potential function is optimized through an iterative training process;

[0041] 4) using the optimized machine learning potential function to predict and optimize the thermodynamic and mechanical properties of multi-principal ceramic materials, including the thermal expansion coefficient, Young's modulus, shear modulus, bulk modulus, yield strength, fracture toughness and laser radiation protection capability of the materials;

[0042] It should be noted that in material design, traditional classical force field methods often have certain accuracy limitations, especially when describing complex multi-principal systems. For this reason, the present invention chooses to use the Deep Potential (DP) model to develop machine learning potential functions. The deep potential model maps atomic positions to energy and force through a deep neural network, thereby providing more accurate and efficient predictions than traditional potential functions. DPGEN can generate training data suitable for machine learning from first-principles calculations and construct machine learning potential functions suitable for complex material systems. The advantage of DPGEN is that it can automatically generate a large amount of high-quality training data, while improving the sampling quality of the data set through an iterative training process.

[0043] The multi-principal-component ceramic material obtained after calculation and optimization using the above method is (Ti0.2Zr0.2Hf0.2Gd0.2Ta0.2)C, which has high yield strength and excellent plastic deformation ability, and can maintain a stable structure under high-energy laser irradiation.

[0044] The machine learning potential function is optimized through an iterative training process that includes an exploration phase, a labeling phase, and a training phase to improve the quality of the training set and the prediction accuracy of the potential function.

[0045] Specifically, in the exploration phase, different configurations of the material are explored through molecular dynamics simulations, covering a variety of possible atomic arrangements and stress states. Each simulation can generate new data points for the expansion of the training set. In this phase, atoms are usually sampled through the "molecular dynamics" method to obtain different configurations and preliminarily calculate their physical quantities such as energy and force.

[0046] Each set of data obtained in the exploration phase needs to be labeled. During the labeling process, the real energy and force obtained by first-principles calculations are used as labels to train the machine learning model. By comparing with the first-principles calculation results, the accuracy of the labeling is ensured, and the data is screened and supplemented based on the differences.

[0047] Based on the labeled data set, the model is trained using a deep neural network. The training process minimizes the loss function (such as energy error, force error, etc.) and optimizes the parameters in the neural network so that the predicted energy and force are as close as possible to the true values ​​calculated by the first principles. Training is usually performed through the back propagation algorithm and the gradient descent method to gradually reduce the error and improve the prediction accuracy of the model.

[0048] The training of machine learning potentials is an iterative process. With each round of training, the training set is expanded based on new simulation results, and the accuracy and efficiency of the training are improved. In each iteration, the model is gradually improved, capturing the interactions between atoms more accurately and accurately predicting more material properties. In this way, the potential function of the machine learning model can gradually adapt to more complex material systems and external environmental conditions.

[0049] The accuracy of the potential function based on machine learning in predicting the properties of multi-principal component ceramic materials was verified by comparing the predicted results of energy and force with the first principles calculation (DFT) results.

[0050] In order to optimize the machine learning potential function, the energy loss function and the force loss function are used in the training process. Specifically, through continuous iterative training, the error between each predicted value and the actual value is optimized, so that the model shows a lower error in the prediction of energy and force. Figure 5 As shown:

[0051] Mean absolute error (MAE) of energy: After comparing the energy prediction results on the test set with the first principle calculation (DFT) results, the mean absolute error (MAE) of energy was 0.02272511eV. This shows that the machine learning model has high accuracy in predicting energy, with a small error range, and can accurately describe the thermodynamic behavior of multi-principal ceramic materials.

[0052] Mean absolute error (MAE) of force: In terms of force prediction, the mean absolute error of the machine learning model is The results show that the machine learning potential function can accurately predict the interaction forces between atoms and can better simulate the mechanical properties of materials under different temperatures and stress states.

[0053] The comparison results of the test set show that the machine learning model can be highly consistent with the results of first-principles calculations in predicting energy and force, such as Figure 6 shown.

[0054] Prediction of thermodynamic and mechanical properties of multi-principal ceramic materials, including the mechanical response of materials under laser radiation, high temperature, compression and tension conditions.

[0055] The Young's modulus of the multi-principal component ceramic material is 349.53 GPa, the shear modulus is 141.94 GPa, the bulk modulus is 216.79 GPa, the Poisson's ratio is 0.23, and its maximum stress is 14.4987 GPa.

[0056] Through the machine learning potential function, the thermal expansion coefficient of the (Ti0.2Zr0.2Hf0.2Gd0.2Ta0.2)C material at temperatures of 400K to 1000K can be predicted. The material has stable thermal expansion behavior at high temperatures.

[0057] Through the developed machine learning potential function, the properties of (Ti0.2Zr0.2Hf0.2Gd0.2Ta0.2)C ceramic materials can be predicted, including their thermodynamic and mechanical behaviors under laser irradiation. The details are as follows:

[0058] Thermodynamic properties prediction:

[0059] The coefficient of thermal expansion (α) is a measure of the volume expansion of a material at different temperatures, usually expressed as the ratio of the length change caused by each unit temperature change. 0.2 Zr 0.2 Hf 0.2 G 0.2 Ta 0.2 )C thermal expansion properties of multi-principal component ceramic materials. This paper predicts the thermal expansion coefficient at different temperatures through molecular dynamics simulation and machine learning model.

[0060] Table 1 Thermal expansion coefficient data of the material at different temperatures (unit: ×10 -6 K -1 )

[0061]

[0062]

[0063] It can be seen from the table that the thermal expansion coefficient of the material shows a certain fluctuation in the temperature range of 400K to 1000K. Specifically, the thermal expansion coefficient of the material at 500K reaches a maximum value of 9.74×10 -6 K -1 , while between 700K and 1000K, the thermal expansion coefficient remains between 8.80 and 9.02×10 -6 K -1 Fluctuates between.

[0064] The data show that (Ti 0.2 Zr 0.2 Hf 0.2 G 0.2 Ta 0.2 )C multi-principal ceramic materials have relatively stable thermal expansion properties at high temperatures, which is of great significance for laser protection applications. Under high temperature environments, the thermal expansion coefficient of the material is low and relatively stable, which helps to reduce material deformation and damage caused by thermal stress, thereby improving the long-term stability and damage resistance of the material under extreme conditions such as laser radiation.

[0065] Mechanical properties prediction:

[0066] The present invention uses DPGEN Auto-Test, an automated tool based on deep learning, to generate and verify machine learning potential functions. In this process, DPGEN generates deep potentials through machine learning methods and applies the potential functions to perform molecular dynamics simulations on materials, further calculating mechanical properties such as bulk modulus, shear modulus, Young's modulus, and Poisson's ratio of the materials.

[0067] (1) The bulk modulus of a material indicates its ability to resist volume changes under uniform pressure. For the (Ti0.2Zr0.2Hf0.2Gd0.2Ta0.2)C material, the calculated bulk modulus is 216.79 GPa. A higher bulk modulus indicates that the material has a strong ability to resist compression and can effectively withstand the thermal expansion and compressive stress caused by laser radiation.

[0068] (2) The shear modulus reflects the deformation ability of a material under shear stress. For this material, the calculated shear modulus is 141.94 GPa. This high shear modulus indicates that the material can maintain good shape stability and deformation resistance when dealing with local stress changes caused by laser radiation.

[0069] (3) Young's modulus describes the rigidity of a material under tension or compression. The Young's modulus of this material is 349.53 GPa, showing its high rigidity under tension and compression. The high Young's modulus indicates that (Ti0.2Zr0.2Hf0.2Gd0.2Ta0.2)C has excellent anti-deformation properties and is particularly suitable for stress concentration environments caused by high-power laser radiation.

[0070] (4) Poisson’s ratio is a ratio that describes the degree of transverse contraction (or expansion) caused by the stretching (or compression) of a material in one direction. The Poisson’s ratio of (Ti0.2Zr0.2Hf0.2Gd0.2Ta0.2)C is 0.23, which means that when the material is subjected to stress, the ratio of transverse to longitudinal deformation is low, and it has good anti-deformation stability.

[0071] (5) When the strain is 0.0739 (i.e. 7.39%), the maximum stress borne by the material is 14.4987 GPa, as Figure 2 This point shows that the material is very strong and can withstand extremely large external forces without breaking. In laser protection applications, materials usually need to withstand temperature changes and thermal stress caused by high-energy laser irradiation, which proves that the material can provide strong bearing capacity in extreme environments; it can continue to deform plastically after the maximum stress point, indicating that it exhibits strong toughness in high-stress environments, and can effectively absorb and disperse local thermal stress caused by laser irradiation, reducing the risk of brittle fracture.

[0072] These mechanical performance parameters show that (Ti0.2Zr0.2Hf0.2Gd0.2Ta0.2)C high entropy ceramic materials can maintain good mechanical stability and damage resistance under extreme environments such as laser radiation and high temperature. The higher bulk modulus and shear modulus ensure that the material can effectively resist pressure under laser radiation, while the higher Young's modulus ensures that the material will not undergo significant deformation under high-power laser. In addition, the moderate Poisson's ratio further improves the mechanical stability of the material.

[0073] Multi-principal ceramic materials can effectively cope with high-power laser irradiation and high-temperature environments, and are suitable for laser weapons, protective equipment, aerospace and nuclear energy fields.

[0074] Therefore, the present invention adopts a design method of a multi-principal ceramic material for laser protection of the above structure, including steps from modeling of the atomic structure of the material to machine learning model training, material performance prediction, etc., to efficiently optimize the thermodynamic and mechanical properties of the (Ti0.2Zr0.2Hf0.2Gd0.2Ta0.2)C ceramic material, and ensure its application under extreme conditions such as laser protection. The obtained (Ti0.2Zr0.2Hf0.2Gd0.2Ta0.2)C high entropy ceramic material can maintain good mechanical stability and damage resistance under extreme environments such as laser radiation and high temperature. The higher bulk modulus and shear modulus ensure that the material can effectively resist pressure under laser radiation, while the higher Young's modulus ensures that the material will not undergo significant deformation under the action of high-power lasers. In addition, the moderate Poisson's ratio further improves the mechanical stability of the material.

[0075] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A design method for multi-principal ceramic materials for laser protection, characterized in that: The method comprises the following steps: 1) The atomic structure of high entropy ceramics is constructed by using the ATAT tool. The structure includes 40 atoms, including titanium, zirconium, hafnium, gadolinium and tantalum elements, and two sets of face-centered cubic structures are nested to obtain a model of multi-principal ceramic materials; 2) Generate stress-strain data, energy and force data of multi-principal ceramic materials based on first-principles calculations, and construct training data sets for machine learning, including tensile, compressive and perturbation configurations; 3) Generate machine learning potential functions using the DPGEN tool, in which atomic positions are mapped to energy and forces through a deep potential model, and the potential function is optimized through an iterative training process; 4) Use the optimized machine learning potential function to predict and optimize the thermodynamic and mechanical properties of multi-principal component ceramic materials, including the thermal expansion coefficient, Young's modulus, shear modulus, bulk modulus, yield strength, fracture toughness and laser radiation protection capability of the material.

2. The design method of a multi-principal-component ceramic material for laser protection according to claim 1, characterized in that: The multi-principal-component ceramic material obtained after calculation and optimization using the above method is (Ti0.2Zr0.2Hf0.2Gd0.2Ta0.2)C, which has high yield strength and excellent plastic deformation ability, and can maintain a stable structure under high-energy laser irradiation.

3. The design method of a multi-principal-component ceramic material for laser protection according to claim 2, characterized in that: The machine learning potential function is optimized through a repeated iterative training process, which includes an exploration phase, a labeling phase, and a training phase to improve the quality of the training set and the prediction accuracy of the potential function.

4. The design method of a multi-principal-component ceramic material for laser protection according to claim 3, characterized in that: The prediction of thermodynamic and mechanical properties of multi-principal ceramic materials also includes the mechanical response of the materials under laser radiation, high temperature, compression and tension conditions.

5. The design method of a multi-principal-component ceramic material for laser protection according to claim 4, characterized in that: The multi-principal component ceramic material has a Young's modulus of 349.53 GPa, a shear modulus of 141.94 GPa, a bulk modulus of 216.79 GPa, a Poisson's ratio of 0.23, and a maximum stress of 14.4987 GPa.

6. The design method of a multi-principal-component ceramic material for laser protection according to claim 5, characterized in that: Through the machine learning potential function, the thermal expansion coefficient of the (Ti0.2Zr0.2Hf0.2Gd0.2Ta0.2)C material at temperatures of 400K to 1000K can be predicted. The material has stable thermal expansion behavior at high temperatures.

7. The design method of a multi-principal-component ceramic material for laser protection according to claim 6, characterized in that: The multi-principal-component ceramic material can effectively cope with high-power laser irradiation and high-temperature environments, and is suitable for laser weapons, protective equipment, aerospace, and nuclear energy fields.

Images