A design method and apparatus for refractory high-entropy alloys based on a large model

Abstract

This invention discloses a design method and apparatus for refractory high-entropy alloys based on a large model, relating to the field of alloy material design technology. The method includes: First, by generating and collecting data, pre-trained surrogate models for thermodynamics (covering phase diagrams, solidification processes, etc.) and mechanical properties are constructed, and calculation models for physical parameters such as valence electron concentration and density are established. Then, these models are integrated into a single ensemble model, and a large language model is introduced as the core of intelligent scheduling. The large model is responsible for parsing the design task, intelligently planning and calling the interfaces of each model based on a knowledge base and rule base. Finally, based on the set design objectives, a multi-objective genetic algorithm is initiated for iterative optimization and adaptive adjustment, efficiently outputting a final alloy design scheme that meets multiple performance requirements. This invention enables intelligent multi-objective design of refractory high-entropy alloys, solving the problem of high dependence on manual intervention.

Description

Technical Field

[0001] This invention relates to the field of alloy material design technology, and in particular to a design method and apparatus for refractory high-entropy alloys based on a large model. Background Technology

[0002] Refractory high-entropy alloys, due to their high melting point, high strength, excellent high-temperature stability, and corrosion resistance, have broad application prospects in aerospace, high-temperature structural components, and nuclear energy. Typical refractory high-entropy alloy systems are usually composed of elements such as Ti, Zr, Hf, Nb, V, Mo, Ta, W, Al, and Cr, with at least five elements involved. Multiple principal element designs are used to achieve synergistic effects of various strengthening mechanisms. As service conditions evolve towards higher temperatures and harsher environments, higher comprehensive requirements are placed on the thermal stability, mechanical properties, and physical parameters of refractory high-entropy alloys, prompting alloy design methods to evolve from traditional empirical trial-and-error approaches to computationally and data-driven approaches. Current refractory high-entropy alloy design primarily relies on numerical calculations and data analysis. On the one hand, CALPHAD-based thermodynamic calculations can utilize commercial or self-built databases to predict phase diagrams, solidification paths, and equilibrium phase compositions for given alloy systems, providing thermodynamic basis for composition screening and microstructure control. On the other hand, researchers also extensively utilize experimental databases, literature data, and empirical formulas to evaluate various empirical parameters such as atomic size mismatch, mixing entropy, and Ω parameter to assist in judging the stability and strengthening potential of alloy systems. Based on this, combining machine learning regression or classification models to predict mechanical properties such as strength and hardness, and using intelligent optimization methods such as genetic algorithms to search in multi-dimensional composition spaces, has become a typical design approach. However, in practical applications, the aforementioned thermodynamic calculations, empirical design, machine learning models, and intelligent optimization algorithms often exist as independent tools, lacking a unified model integration and collaboration mechanism. Researchers need to frequently switch between different software and scripts and manually transfer data. The multi-objective optimization process heavily relies on personal experience and is not user-friendly for non-algorithm experts. As design requirements rapidly evolve from single performance indicators to multi-objective comprehensive performance balance, single-model or fixed-configuration optimization algorithms struggle to efficiently obtain reasonable Pareto compromise solutions under high-dimensional composition spaces and complex constraints. Meanwhile, although large-scale language models have demonstrated outstanding capabilities in natural language understanding, task planning, and code generation, there is still a lack of systematic technical solutions on how to deeply collaborate with thermodynamic proxy models, mechanical property proxy models, and physical parameter calculation modules to build an intelligent design framework for refractory high-entropy alloys. Summary of the Invention

[0003] To address the technical problems of existing technologies, such as fragmented models, low efficiency in multi-objective optimization, and high dependence on manual processes, this invention provides a design method and apparatus for refractory high-entropy alloys based on a large model. The technical solution is as follows:

[0004] On the one hand, a design method for refractory high-entropy alloys based on a large model is provided. This method is implemented by a design device for refractory high-entropy alloys based on a large model, and includes: S1: Determine the composition space and thermodynamic content of the refractory high-entropy alloy system, generate a thermodynamic dataset, perform regression modeling on the thermodynamic response based on the thermodynamic dataset, and obtain a pre-trained thermodynamic proxy model and thermodynamic interaction interface. The thermodynamic content includes phase diagram, Scheil solidification process and equilibrium phase composition. S2: Obtain experimental data and feature sets describing the mechanical properties of refractory high-entropy alloys, train them, and obtain a pre-trained mechanical property prediction proxy model. S3: Construct a computational model for the target physical parameters, including valence electron concentration, atomic size mismatch, mixing entropy, and density; S4: Construct an integrated model and an integrated interface. Based on the integrated model and the integrated interface, construct a scheduler. The construction of the scheduler includes accepting design tasks and sending them to the large model. The large model analyzes the requirements of the design tasks and determines the internal calling order and optimization scheme of the integrated model and the integrated interface. The large model analyzes the requirements of the design tasks and sets calling principles based on the integrated model knowledge base and the rule base of the integrated interface, and generates corresponding prompt words according to the calling principles. S5: Construct a design task, set the genetic algorithm parameters based on the design task, and construct a multi-objective genetic algorithm; S6: Based on a multi-objective genetic algorithm, perform multi-objective optimization iterations and adaptively adjust the optimization strategy to obtain the final alloy design result.

[0005] Preferably, in step S1, the composition space and thermodynamic content of the refractory high-entropy alloy system are determined, a thermodynamic dataset is generated, and regression modeling of the thermodynamic response is performed based on the thermodynamic dataset to obtain a pre-trained thermodynamic surrogate model and a thermodynamic interaction interface. The thermodynamic content includes a phase diagram, the Scheil solidification process, and the equilibrium phase composition, including: S11: Determine the composition space of the refractory high-entropy alloy system, wherein the composition space includes the types of elements, the content range of each element, and the calculation step size; S12: Determine the thermodynamic content to be calculated based on the application objectives. The thermodynamic content includes phase diagrams, Scheil solidification process, and equilibrium phase composition. S13: Based on the composition space and the thermodynamic content, use high-throughput thermodynamic calculation methods to perform batch calculations to obtain thermodynamic data; S14: Establish a structured specification for thermodynamic data, analyze and extract the thermodynamic data, and perform structured integration to obtain a thermodynamic dataset; S15: Regression modeling of thermodynamic response is performed based on thermodynamic dataset to obtain a pre-trained thermodynamic surrogate model. The regression modeling includes modeling using one or more machine learning models such as random forest, gradient boosting tree, support vector machine, decision tree or neural network. S16: Based on the thermodynamic data structure specifications, establish a thermodynamic interaction interface. The thermodynamic interaction interface is used to input parameters to the thermodynamic proxy model based on the large model and obtain the corresponding thermodynamic prediction results.

[0006] Preferably, step S2 involves acquiring experimental data and feature sets describing the mechanical properties of refractory high-entropy alloys, training them, and obtaining a pre-trained mechanical property prediction proxy model, including: S21: Obtain experimental data of refractory high-entropy alloys, wherein obtaining experimental data of refractory high-entropy alloys includes extracting and organizing data related to the target system and target mechanical property indicators from refractory high-entropy alloy related databases, published literature and proprietary experimental data; S22: Establish a structured specification for mechanical property data, and obtain a feature set for describing mechanical properties. The acquisition of the feature set for describing mechanical properties includes constructing the feature set based on materials science theory, combined with elemental characteristics, thermodynamic characteristics, phase composition characteristics, heat treatment process parameters and microstructure characteristics, and using feature engineering methods. S23: Based on the feature set and experimental data, training is performed to obtain a pre-trained mechanical performance prediction proxy model. The training includes using one or more machine learning models such as random forest, gradient boosting tree, support vector machine, decision tree or neural network, with the feature set as input and the experimental data as output. S24: Based on the structured specifications of mechanical performance data, establish a mechanical performance data interaction interface. The mechanical performance data interaction interface is used to input parameters into the thermodynamic performance prediction proxy model based on the large model and obtain the corresponding mechanical performance prediction results.

[0007] Preferably, the calculation model for constructing the target physical parameters in S3 includes valence electron concentration, atomic size mismatch, mixing entropy, and density, comprising: S31: Construct a calculation model for the target physical parameters. The calculation model for the target physical parameters includes constructing a calculation model for the target physical parameters based on the theory of refractory high-entropy alloys and related physical models. The target physical parameters include valence electron concentration, atomic size mismatch, mixing entropy, and density. S32: Construct a structured specification for the target physical parameter data; S33: Based on the structured specifications of the target physical parameter data, establish a target physical parameter data interaction interface. The target physical parameter data interaction interface is used to input the target physical parameter data into the calculation model of the target physical parameter based on the large model and obtain the corresponding physical parameter calculation results.

[0008] Preferably, in step S4, the construction of an integration model and an integration interface, and the construction of a scheduler based on the integration model and the integration interface, wherein the scheduler includes accepting design tasks and sending them to a large model, the large model determining the internal calling order and optimization scheme of the integration model and the integration interface according to the design task analysis requirements, and the large model, based on the design task analysis requirements, setting calling principles according to the integration model knowledge base and the integration interface rule base, and generating corresponding prompts according to the calling principles, including: S41: Integrate the thermodynamic proxy model, the mechanical property prediction proxy model, and the calculation model of the target physical parameters to construct an integrated model. At the same time, construct an integrated model knowledge base based on the thermodynamic proxy model, the mechanical property prediction proxy model, and the calculation model of the target physical parameters. The integrated model knowledge base limits the model parameters and simulation space in combination with the characteristics of the refractory high-entropy alloy system. S42: Integrate the thermodynamic interaction interface, the mechanical performance data interaction interface, and the target physical parameter data interaction interface to construct an integrated interface; S43: Construct a scheduler to manage the calls to the integrated model and the integrated interface. The scheduler includes accepting design tasks and sending them to the large model. The large model analyzes the requirements of the design tasks, determines the internal calling order of the integrated model and the integrated interface, and the optimization scheme. The design tasks include the composition design range, constraints, and multi-objective tasks of the refractory high-entropy alloy to be designed. The optimization scheme includes determining the objective function, decision variable representation, constraints, and genetic operators of the multi-objective optimization problem. The large model analyzes the requirements of the design tasks, including setting calling principles based on the integrated model knowledge base and the rule base of the integrated interface, and generating corresponding prompts according to the calling principles.

[0009] Preferably, the construction design task of S5, based on the design task, sets the genetic algorithm parameters and constructs a multi-objective genetic algorithm, including: S51: Based on the composition design range, constraints and multi-objective tasks of the refractory high-entropy alloy to be designed, a design task is constructed. The multi-objective tasks include quantifiable performance indicators of the refractory high-entropy alloy to be designed under target service conditions. The performance indicators include at least one or more of thermal stability-related indicators, strength and hardness-related mechanical property indicators and density-related physical parameters. The design task defines the composition design space. S52: Set the genetic algorithm parameters and construct a multi-objective genetic algorithm for searching the Pareto front of the target performance.

[0010] Preferably, step S6, based on a multi-objective genetic algorithm, performs multi-objective optimization iterations and adaptively adjusts the optimization strategy to obtain the final alloy design result, including: S61: Based on the design task, the multi-objective genetic algorithm is invoked to randomly initialize N individuals within the given composition design space, where each individual represents a candidate alloy composition scheme; S62: Input each individual into the ensemble model through the scheduler to obtain the multidimensional performance index of each individual; S63: Evaluate the merits and demerits of each individual based on the aforementioned multidimensional performance indicators, and determine the direction of evolution; S64: Select parent individuals from the current population, perform crossover and mutation operations to generate N new offspring individuals; S65: Through multiple generations of iterative evolution, the population is continuously updated until a Pareto front solution set that approximates the target performance is obtained; S66: Analyze the Pareto front solution set obtained based on the large model. If the design goal or constraints are not met, adaptively adjust the design task and genetic algorithm parameters until the preset convergence condition is met or the predetermined number of iterations is reached to obtain the optimal design task. S67: Combine the optimal design task with the corresponding multi-objective performance prediction value and the difference or deviation index from the target performance to obtain the final alloy design result.

[0011] On the other hand, a design apparatus for refractory high-entropy alloys based on a large model is provided. This apparatus is applied to a design method for refractory high-entropy alloys based on a large model. The apparatus includes: Thermodynamics module: used to determine the composition space and thermodynamic content of refractory high-entropy alloy systems, generate thermodynamic datasets, perform regression modeling on thermodynamic responses based on thermodynamic datasets, and obtain pre-trained thermodynamic surrogate models and thermodynamic interaction interfaces. The thermodynamic content includes phase diagrams, Scheil solidification process, and equilibrium phase composition. Mechanical properties module: used to acquire experimental data and feature sets describing the mechanical properties of refractory high-entropy alloys, train them, and obtain a pre-trained mechanical property prediction surrogate model; Physical Parameters Module: Used to construct computational models for target physical parameters, including valence electron concentration, atomic size mismatch, mixing entropy, and density; Integrated Model Module: Used to build an integrated model, build an integrated interface, and build a scheduler based on the integrated model and the integrated interface. The scheduler construction includes accepting design tasks and sending them to the large model. The large model analyzes the requirements of the design task and determines the internal calling order and optimization scheme of the integrated model and the integrated interface. The large model analyzes the requirements of the design task and sets calling principles based on the integrated model knowledge base and the rule base of the integrated interface, and generates corresponding prompt words according to the calling principles. Multi-objective genetic algorithm module: used to construct design tasks, set genetic algorithm parameters based on design tasks, and construct multi-objective genetic algorithms; Optimization module: Used to perform multi-objective optimization iterations based on multi-objective genetic algorithms and adaptively adjust the optimization strategy to obtain the final alloy design result.

[0012] On the other hand, a design apparatus for refractory high-entropy alloys based on a large model is provided. The design apparatus for refractory high-entropy alloys based on a large model includes: a processor; and a memory storing computer-readable instructions. When the computer-readable instructions are executed by the processor, they implement the method described in any of the above-described design methods for refractory high-entropy alloys based on a large model.

[0013] On the other hand, a computer-readable storage medium is provided, characterized in that the computer-readable storage medium stores program code, which can be invoked by a processor to execute the method as described in any one of claims 1 to 7.

[0014] The beneficial effects of the technical solutions provided by the embodiments of the present invention include at least the following: This invention addresses the problems of model fragmentation, low efficiency of multi-objective optimization, and high dependence on manual intervention in the intelligent multi-objective design of refractory high-entropy alloys. Compared with comparative methods relying on traditional machine learning or simple thermodynamic calculations, the candidate alloys obtained in the embodiments of this invention exhibit superior performance in balancing multi-objective properties, and the model invocation and search steps are simpler and the search efficiency is higher. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart of a design method for refractory high-entropy alloys based on a large model, provided by an embodiment of the present invention; Figure 2 This is a schematic diagram of the Pareto front (2-1) and room temperature compressive mechanical properties (2-2) of a refractory high-entropy alloy provided in an embodiment of the present invention; Figure 3 This is a block diagram of a design device for refractory high-entropy alloys based on a large model, provided in an embodiment of the present invention. Figure 4 This is a schematic diagram of the structure of a design device for refractory high-entropy alloys based on a large model, provided in an embodiment of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0018] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0019] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0020] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0021] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0022] This invention provides a design method for refractory high-entropy alloys based on a large model. This method can be implemented using a large-model-based refractory high-entropy alloy design device, which can be a terminal or a server. Figure 1 The flowchart shown is for a design method of refractory high-entropy alloys based on a large model. The processing flow of this method may include the following steps:

[0023] The composition space and thermodynamic content of the refractory high-entropy alloy system are determined, a thermodynamic dataset is generated, and regression modeling of the thermodynamic response is performed based on the thermodynamic dataset to obtain a pre-trained thermodynamic surrogate model and a thermodynamic interaction interface. The thermodynamic content includes phase diagram, Scheil solidification process and equilibrium phase composition. Preferably, the composition space and thermodynamic content of the refractory high-entropy alloy system are determined, a thermodynamic dataset is generated, and regression modeling of the thermodynamic response is performed based on the thermodynamic dataset to obtain a pre-trained thermodynamic surrogate model and a thermodynamic interaction interface. The thermodynamic content includes phase diagrams, the Scheil solidification process, and equilibrium phase composition, including: Determine the composition space of the refractory high-entropy alloy system, wherein the composition space includes the types of elements, the content range of each element, and the calculation step size; The thermodynamic content to be calculated is determined based on the application objectives. The thermodynamic content includes phase diagrams, Scheil solidification process, and equilibrium phase composition. Based on the composition space and the thermodynamic content, high-throughput thermodynamic calculation methods are used to perform batch calculations to obtain thermodynamic data; A thermodynamic data structure specification is established, and the thermodynamic data is parsed and extracted, and then structured and integrated to obtain a thermodynamic dataset. Regression modeling of thermodynamic response is performed based on thermodynamic dataset to obtain a pre-trained thermodynamic surrogate model. The regression modeling includes modeling using one or more machine learning models such as random forest, gradient boosting tree, support vector machine, decision tree or neural network. Based on the aforementioned thermodynamic data structure specifications, a thermodynamic interaction interface is established. This interface is used to input parameters into the thermodynamic proxy model based on the large model and obtain the corresponding thermodynamic prediction results.

[0024] In some embodiments, in order to obtain a refractory high-entropy alloy with both high room temperature strength and high plasticity, for a six-element refractory high-entropy alloy of Ti, Zr, Hf, Nb, Mo, and Ta, by controlling the content of each element, under the premise of meeting the service temperature of 1000-1200 ℃, density less than 9 g / cm³, and melting point as high as possible, a multi-objective synergistic optimization is achieved to achieve a room temperature compressive yield strength greater than 950 MPa and a room temperature fracture strain greater than 45%.

[0025] It should be noted that high-throughput thermodynamic calculations are performed to obtain equilibrium phase diagrams at different temperatures, and the melting point, liquidus line, and phase content of each alloy are quickly extracted to form a dataset.

[0026] It should be further explained that the mole fraction range of six elements, Ti, Zr, Hf, Nb, Mo, and Ta, can be selected from 5% to 35%. Representative composition points are generated with a step size of % while ensuring that the sum of the mole fractions of each element is 100%. Equilibrium phase diagrams are calculated for each composition point at different temperatures. Key data such as solidus temperature, liquidus temperature, and main phase volume fraction are extracted. A deep learning model is used to establish the mapping relationship from composition and temperature to thermodynamic response, thus forming a thermodynamic proxy model.

[0027] We acquire experimental data and feature sets describing the mechanical properties of refractory high-entropy alloys, train them, and obtain a pre-trained surrogate model for predicting mechanical properties. Preferably, experimental data and feature sets describing the mechanical properties of refractory high-entropy alloys are acquired, and trained to obtain a pre-trained surrogate model for predicting mechanical properties, including: The experimental data for obtaining refractory high-entropy alloys includes extracting and organizing data related to the target system and target mechanical property indicators from refractory high-entropy alloy related databases, published literature and proprietary experimental data. Establish a structured specification for mechanical property data, and obtain a feature set for describing mechanical properties. The acquisition of the feature set for describing mechanical properties includes constructing the feature set based on materials science theory, combined with elemental characteristics, thermodynamic characteristics, phase composition characteristics, heat treatment process parameters and microstructure characteristics, and using feature engineering methods. Based on the feature set and experimental data, a pre-trained mechanical performance prediction proxy model is obtained through training. The training includes using one or more machine learning models such as random forest, gradient boosting tree, support vector machine, decision tree or neural network, with the feature set as input and the experimental data as output. Based on the structured specifications of mechanical property data, a mechanical property data interaction interface is established. The mechanical property data interaction interface is used to input parameters into the thermodynamic property prediction proxy model based on the large model and obtain the corresponding mechanical property prediction results.

[0028] In some embodiments, based on actual application requirements, experimental data on room temperature compressive yield strength and room temperature fracture strain of TiZrHfNbMoTa and related refractory high-entropy alloys are collected. Based on materials science theory and feature engineering, a feature set describing the strength and plasticity of the alloy is constructed. The mechanical property prediction proxy model of room temperature compressive yield strength and room temperature fracture strain is obtained by training the feature set.

[0029] It should be noted that, based on the collected literature and experimental data, a comprehensive feature set including elemental physicochemical parameters, mixing entropy, atomic size mismatch, phase composition characteristics, and heat treatment parameters can be constructed. The room temperature compressive yield strength and room temperature fracture strain are used as outputs, respectively. Cross-validation is used to select the model and feature subset with smaller prediction error as a proxy model for mechanical property prediction, which is then used for rapid evaluation of strength and plasticity indices in subsequent multi-objective genetic algorithms.

[0030] A computational model for the target physical parameters is constructed, including valence electron concentration, atomic size mismatch, mixing entropy, and density. Preferably, a computational model for the target physical parameters is constructed, wherein the target physical parameters include valence electron concentration, atomic size mismatch, mixing entropy, and density, including: A calculation model for the target physical parameters is constructed. The calculation model for the target physical parameters includes valence electron concentration, atomic size mismatch, mixing entropy, and density, based on the theory of refractory high-entropy alloys and related physical models. Construct a structured specification for the target physical parameter data; Based on the structured specifications of the target physical parameter data, a target physical parameter data interaction interface is established. The target physical parameter data interaction interface is used to input the target physical parameter data into the calculation model of the target physical parameter based on the large model and obtain the corresponding physical parameter calculation results.

[0031] In some embodiments, a calculation script for calculating density is written based on the theory of refractory high-entropy alloys, forming a calculation model that can output corresponding target physical parameters according to the alloy composition.

[0032] It should be noted that the theoretical density of the candidate components is estimated by the physical parameter calculation module, and alloy component combinations with a density greater than 9 g / cm³ are eliminated in advance to constrain the candidate component space.

[0033] An integrated model and an integrated interface are constructed. A scheduler is constructed based on the integrated model and the integrated interface. The construction of the scheduler includes accepting design tasks and sending them to a large model. The large model analyzes the requirements of the design tasks and determines the internal calling order and optimization scheme of the integrated model and the integrated interface. The large model analyzes the requirements of the design tasks and sets calling principles based on the integrated model knowledge base and the rule base of the integrated interface, and generates corresponding prompt words according to the calling principles. Preferably, an integrated model and an integrated interface are constructed. A scheduler is then constructed based on the integrated model and the integrated interface. The construction of the scheduler includes accepting design tasks and sending them to a large model. The large model analyzes the requirements of the design tasks and determines the internal calling order and optimization scheme of the integrated model and the integrated interface. The large model's analysis of the design tasks includes setting calling principles based on the integrated model's knowledge base and the integrated interface's rule base, and generating corresponding prompts according to these principles, including: The thermodynamic proxy model, the mechanical property prediction proxy model, and the calculation model of the target physical parameters are integrated to construct an integrated model. At the same time, an integrated model knowledge base is constructed based on the thermodynamic proxy model, the mechanical property prediction proxy model, and the calculation model of the target physical parameters. The integrated model knowledge base limits the model parameters and simulation space in combination with the characteristics of the refractory high-entropy alloy system. The thermodynamic interaction interface, the mechanical performance data interaction interface, and the target physical parameter data interaction interface are integrated to construct an integrated interface; A scheduler is constructed to manage the invocation of the integrated model and the integrated interface. The scheduler includes accepting design tasks and sending them to the large model. The large model analyzes the requirements of the design tasks, determines the internal invocation order of the integrated model and the integrated interface, and the optimization scheme. The design tasks include the composition design range, constraints, and multi-objective tasks of the refractory high-entropy alloy to be designed. The optimization scheme includes determining the objective function, decision variable representation, constraints, and genetic operators of the multi-objective optimization problem. The large model analyzes the requirements of the design tasks, including setting invocation principles based on the integrated model knowledge base and the rule base of the integrated interface, and generating corresponding prompts according to the invocation principles.

[0034] In some embodiments, by interactively inputting the composition design range of the TiZrHfNbMoTa hexa-element refractory high-entropy alloy into a large model, a multi-objective design task and constraints are given, including a service temperature of 1000–1200 °C, a density of less than 9 g / cm³, a melting point and phase content as high as possible, a room temperature compressive yield strength greater than 950 MPa, and a room temperature fracture strain greater than 45%. The large model parses the objectives and constraints in natural language form, constructs a multi-objective genetic algorithm with alloy composition as the decision variable, and completes population initialization.

[0035] It should be noted that the large model calls the thermodynamic proxy model, the mechanical property prediction proxy model, and the density calculation module during the multi-objective genetic algorithm iteration process to predict and evaluate the phase content, density, solidus temperature, room temperature compressive yield strength, and room temperature fracture strain properties of each candidate TiZrHfNbMoTa alloy composition in the population. It performs multiple generations of evolution according to the selection, crossover, and mutation rules of the multi-objective genetic algorithm to obtain the Pareto front solution set that approximates the target performance under the premise of satisfying the density and service temperature constraints, and selects and outputs several candidate alloy design schemes from them.

[0036] It should be further explained that the large model automatically converts the input multi-objective tasks such as "density less than 9 g / cm³", "solidspot temperature as high as possible", "room temperature compressive yield strength greater than 950 MPa", and "room temperature fracture strain greater than 45%" into objective functions and constraints in a multi-objective optimization problem. The mole fractions of Ti, Zr, Hf, Nb, Mo, and Ta are used as decision variables. In each generation of the genetic algorithm evolution, the thermodynamic surrogate model, the mechanical property prediction surrogate model, and the physical parameter calculation module are called to evaluate the fitness of individuals. Through multiple generations of iteration, a Pareto front solution set that meets the comprehensive requirements of thermal stability, strength, plasticity, and density is obtained.

[0037] Construct a design task, set the genetic algorithm parameters based on the design task, and construct a multi-objective genetic algorithm; Preferably, a design task is constructed, genetic algorithm parameters are set based on the design task, and a multi-objective genetic algorithm is constructed, including: Based on the compositional design range, constraints, and multi-objective tasks of the refractory high-entropy alloy to be designed, a design task is constructed. The multi-objective tasks include quantifiable performance indicators of the refractory high-entropy alloy to be designed under target service conditions. The performance indicators include at least one or more of thermal stability-related indicators, strength and hardness-related mechanical property indicators, and density-related physical parameters. The design task defines the compositional design space. Set the parameters of the genetic algorithm and construct a multi-objective genetic algorithm for searching the Pareto front of the target performance.

[0038] Based on a multi-objective genetic algorithm, multi-objective optimization iterations are performed and the optimization strategy is adaptively adjusted to obtain the final alloy design result.

[0039] Preferably, based on a multi-objective genetic algorithm, multi-objective optimization iterations are performed and the optimization strategy is adaptively adjusted to obtain the final alloy design result, including: Based on the design task, the multi-objective genetic algorithm is invoked to randomly initialize N individuals within a given composition design space, where each individual represents a candidate alloy composition scheme. Each individual is input into the ensemble model through a scheduler to obtain multidimensional performance indicators for each individual; The merits and demerits of each individual are evaluated based on the aforementioned multidimensional performance indicators to determine the direction of evolution; Select parent individuals from the current population, perform crossover and mutation operations to generate N new offspring individuals; Through multiple generations of iterative evolution, the population is continuously updated until a Pareto front solution set that approximates the target performance is obtained; Based on the large model, the Pareto front solution set is analyzed. If the design goal or constraints are not met, the design task and genetic algorithm parameters are adaptively adjusted until the preset convergence condition is met or the predetermined number of iterations is reached, so as to obtain the optimal design task. The final alloy design result is obtained by combining the optimal design task with the corresponding multi-objective performance prediction value and the difference or deviation index from the target performance.

[0040] In some embodiments, the candidate refractory high-entropy alloy composition of Ti30Zr15Hf18Nb30Mo4Ta3 output by the large model is prepared by vacuum arc melting and tested for room temperature compressive properties. The alloy density, room temperature compressive yield strength and room temperature fracture strain are measured. The test results are compared with the design targets of density less than 9 g / cm³, room temperature compressive yield strength greater than 950 MPa and room temperature fracture strain greater than 45%. When the test data meets the above multi-objective performance requirements, the final alloy design scheme is determined. If necessary, the test data is returned to update the surrogate model and the large model drives the repeated iterative optimization process.

[0041] It should be noted that, such as Figure 2 As shown, through the intelligent design method based on large model collaboration of this invention, multiple sets of TiZrHfNbMoTa refractory high entropy alloy candidate compositions that meet the comprehensive index of "service temperature of 1000-1200 ℃, density of less than 9 g / cm³, room temperature compressive yield strength of greater than 950MPa, and room temperature fracture strain of greater than 45%" can be obtained.

[0042] Another embodiment is provided below: This embodiment focuses on a six-element refractory high-entropy alloy of Ti, Zr, Hf, Nb, Mo, and Ta. By controlling the content of each element, it achieves multi-objective synergistic optimization of room temperature compressive yield strength and room temperature fracture strain under the premise of density less than 8.5 g / cm³ and single phase at 1000℃. The composition range and calculation step size of each element are determined, and high-throughput thermodynamic calculations are performed to obtain the equilibrium phase diagram at 1000℃. The content of each alloy phase is quickly extracted to form a dataset, and a thermodynamic surrogate model for rapid prediction of thermodynamic response is established based on the thermodynamic data. By interacting with a large model and inputting the composition design range of the TiZrHfNbMoTa six-element refractory high-entropy alloy, a multi-objective design task and constraints are given for density less than 9 g / cm³, single phase at 1000℃, and room temperature strong-plastic equilibrium. The large model parses the objectives and constraints in natural language form, constructs a multi-objective genetic algorithm with alloy composition as the decision variable, and completes population initialization.

[0043] The large model calls the thermodynamic proxy model, the mechanical property prediction proxy model, and the density calculation module during the multi-objective genetic algorithm iteration process to predict and evaluate the phase content, density, room temperature compressive yield strength, and room temperature fracture strain properties of each candidate TiZrHfNbMoTa alloy composition in the population. According to the selection, crossover, and mutation rules of the multi-objective genetic algorithm, multiple generations of evolution are carried out to obtain the Pareto front solution set that approximates the target performance under the premise of satisfying the density and service temperature constraints. From this set, several candidate alloy design schemes are selected and output.

[0044] The candidate refractory high-entropy alloy Ti35Zr29Hf4Nb20Mo5Ta5 output by the large model was prepared by vacuum arc melting and tested for room temperature compressive properties. The alloy density, room temperature compressive yield strength and room temperature fracture strain were measured. The test results were compared with the design objectives. When the test data met the above multi-objective performance requirements, the final alloy design scheme was determined. If necessary, the test data was returned to update the surrogate model and the large model drove the repeated iterative optimization process.

[0045] In the embodiments, the intelligent design method based on large model collaboration of the present invention can be used to obtain multiple sets of single-phase TiZrHfNbMoTa refractory high-entropy alloys with a density of less than 8.5 g / cm³, a room temperature compressive yield strength of 1300 MPa, and a room temperature fracture strain of 35%.

[0046] Another embodiment is provided below: This embodiment focuses on a seven-element refractory high-entropy alloy of Cr, Ti, Zr, Hf, Nb, Mo, and Ta. By controlling the content of each element, it achieves multi-objective synergistic optimization while satisfying a density of less than 8.9 g / cm³, room temperature compressive plasticity of >25%, higher high-temperature strength, and a smaller Scheil solidification range.

[0047] Taking the CrTiZrHfNbMoTa heptogenic refractory high-entropy alloy system as the object, the composition range and calculation step size of each element were determined, and high-throughput thermodynamic calculations were performed on the system to obtain the Scheil solidification phase diagram of each alloy. The solidification interval of each alloy was quickly extracted, and a thermodynamic surrogate model for rapid prediction of thermodynamic response was established based on the thermodynamic data.

[0048] Experimental data on high-temperature compressive yield strength and room-temperature compressive plasticity of CrTiZrHfNbMoTa and related refractory high-entropy alloys were collected. Based on materials science theory and feature engineering, a feature set describing the high-temperature strength of the alloys was constructed. The mechanical property prediction surrogate model of high-temperature compressive yield strength and room-temperature compressive plasticity was obtained by training the feature set.

[0049] By interacting with a large model and inputting the compositional design range of the CrTiZrHfNbMoTa heptagonal refractory high-entropy alloy, a multi-objective design task and constraints are given, which have a density of less than 8.9 g / cm³ and a room temperature compressive plasticity of >25%. This achieves higher high-temperature strength and a smaller Scheil solidification range. The large model parses the objectives and constraints in natural language form, constructs a multi-objective genetic algorithm with alloy composition as the decision variable, and completes population initialization.

[0050] The large model calls the thermodynamic proxy model, the mechanical property prediction proxy model, and the density calculation module during the multi-objective genetic algorithm iteration process to predict and evaluate the solidification range, density, high-temperature compressive yield strength, and room-temperature plasticity of each candidate alloy composition in the population. According to the selection, crossover, and mutation rules of the multi-objective genetic algorithm, multiple generations of evolution are carried out to obtain the Pareto front solution set that approximates the target performance under the premise of satisfying the density and service temperature constraints. From this set, several candidate alloy design schemes are selected and output.

[0051] The candidate refractory high-entropy alloy composition of Cr13.4Zr0.4Hf0.8Nb26.7Mo28.7Ta30 output by the large model was prepared by vacuum arc melting and tested for room temperature compressive properties. The alloy density, high temperature compressive yield strength and room temperature fracture strain were measured. The test results were compared with the design objectives. When the test data met the above multi-objective performance requirements, the final alloy design scheme was determined. If necessary, the test data was returned to update the surrogate model and the large model drove the repeated iterative optimization process.

[0052] By using a large-model-based intelligent design method, multiple sets of refractory high-entropy alloys can be obtained that meet the following requirements: density less than 8.5 g / cm³, high-temperature compressive yield strength of 645 MPa at 1000℃, and room-temperature fracture strain of 27%.

[0053] The above is an introduction to the method embodiments. The following describes the solution described in this application through device embodiments.

[0054] Figure 3 This is a block diagram illustrating a large-model-based design apparatus for refractory high-entropy alloys, according to an exemplary embodiment. The apparatus is used in a large-model-based design method for refractory high-entropy alloys. (Refer to...) Figure 3 The device includes a thermodynamic module, a mechanical performance module, a physical parameter module, an integrated model module, a multi-objective genetic algorithm module, and an optimization module.

[0055] Thermodynamics module: used to determine the composition space and thermodynamic content of refractory high-entropy alloy systems, generate thermodynamic datasets, perform regression modeling on thermodynamic responses based on thermodynamic datasets, and obtain pre-trained thermodynamic surrogate models and thermodynamic interaction interfaces. The thermodynamic content includes phase diagrams, Scheil solidification process, and equilibrium phase composition. Mechanical properties module: used to acquire experimental data and feature sets describing the mechanical properties of refractory high-entropy alloys, train them, and obtain a pre-trained mechanical property prediction surrogate model; Physical Parameters Module: Used to construct computational models for target physical parameters, including valence electron concentration, atomic size mismatch, mixing entropy, and density; Integrated Model Module: Used to build an integrated model, build an integrated interface, and build a scheduler based on the integrated model and the integrated interface. The scheduler construction includes accepting design tasks and sending them to the large model. The large model analyzes the requirements of the design task and determines the internal calling order and optimization scheme of the integrated model and the integrated interface. The large model analyzes the requirements of the design task and sets calling principles based on the integrated model knowledge base and the rule base of the integrated interface, and generates corresponding prompt words according to the calling principles. Multi-objective genetic algorithm module: used to construct design tasks, set genetic algorithm parameters based on design tasks, and construct multi-objective genetic algorithms; Optimization module: Used to perform multi-objective optimization iterations based on multi-objective genetic algorithms and adaptively adjust the optimization strategy to obtain the final alloy design result.

[0056] A large-model-based refractory high-entropy alloy design device includes: a processor; and a memory storing computer-readable instructions, which, when executed by the processor, implement the method described in any of the above-described large-model-based refractory high-entropy alloy design methods.

[0057] A computer-readable storage medium, characterized in that the computer-readable storage medium stores program code, the program code being invoked by a processor to execute the method as described in any one of claims 1 to 7.

[0058] Figure 4 This is a schematic diagram of a design device for refractory high-entropy alloys based on a large model, provided in an embodiment of the present invention. Figure 4 As shown, the design equipment for refractory high-entropy alloys based on a large model can include the above-mentioned Figure 3 The illustrated apparatus is a large-model-based design device for refractory high-entropy alloys. Optionally, the large-model-based design device 410 for refractory high-entropy alloys may include a first processor 2001.

[0059] Optionally, the refractory high-entropy alloy design device 410 based on a large model may also include a memory 2002 and a transceiver 2003.

[0060] The first processor 2001, memory 2002, and transceiver 2003 can be connected via a communication bus.

[0061] The following is combined with Figure 4 A detailed description of each component of the refractory high-entropy alloy design equipment 410 based on a large model is provided below: The first processor 2001 is the control center of the large-scale refractory high-entropy alloy design device 410. It can be a single processor or a collective term for multiple processing elements. For example, the first processor 2001 can be one or more central processing units (CPUs), application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).

[0062] Optionally, the first processor 2001 can perform various functions of the refractory high-entropy alloy design device 410 based on a large model by running or executing software programs stored in the memory 2002 and calling data stored in the memory 2002.

[0063] In a specific implementation, as one example, the first processor 2001 may include one or more CPUs, for example... Figure 4 CPU0 and CPU1 are shown in the diagram.

[0064] In a specific implementation, as one example, the refractory high-entropy alloy design device 410 based on a large model may also include multiple processors, for example... Figure 4 The first processor 2001 and the second processor 2004 are shown in the diagram. Each of these processors can be a single-core processor or a multi-core processor. Here, a processor can refer to one or more devices, circuits, and / or processing cores used to process data (such as computer program instructions).

[0065] The memory 2002 is used to store the software program that executes the present invention, and is controlled by the first processor 2001 to execute it. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.

[0066] Optionally, the memory 2002 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 2002 may be integrated with the first processor 2001 or may exist independently, and may be connected via the interface circuit of the large-model refractory high-entropy alloy design device 410. Figure 4 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0067] The transceiver 2003 is used to communicate with network devices or with terminal devices.

[0068] Alternatively, transceiver 2003 may include a receiver and a transmitter. Figure 4 (Not shown separately). The receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function.

[0069] Optionally, the transceiver 2003 can be integrated with the first processor 2001, or it can exist independently, and its interface circuit can be used with the large-model refractory high-entropy alloy design device 410. Figure 4 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0070] It should be noted that, Figure 4 The structure of the refractory high-entropy alloy design device 410 shown in the figure does not constitute a limitation on the router. The actual knowledge structure identification device may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0071] Furthermore, the technical effects of the refractory high-entropy alloy design equipment 410 based on a large model can be referred to the technical effects of the refractory high-entropy alloy design method based on a large model described in the above method embodiments, and will not be repeated here.

[0072] It should be understood that the first processor 2001 in the embodiments of the present invention may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0073] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0074] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0075] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0076] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0077] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0078] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0079] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0080] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0081] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0082] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0083] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0084] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A design method for refractory high-entropy alloys based on a large model, characterized in that, The method includes: S1: Determine the composition space and thermodynamic content of the refractory high-entropy alloy system, generate a thermodynamic dataset, perform regression modeling on the thermodynamic response based on the thermodynamic dataset, and obtain a pre-trained thermodynamic proxy model and thermodynamic interaction interface. The thermodynamic content includes phase diagram, Scheil solidification process and equilibrium phase composition. S2: Obtain experimental data and feature sets describing the mechanical properties of refractory high-entropy alloys, train them, and obtain a pre-trained mechanical property prediction proxy model. S3: Construct a computational model for the target physical parameters, including valence electron concentration, atomic size mismatch, mixing entropy, and density; S4: Construct an integrated model and an integrated interface. Based on the integrated model and the integrated interface, construct a scheduler. The construction of the scheduler includes accepting design tasks and sending them to the large model. The large model analyzes the requirements of the design tasks and determines the internal calling order and optimization scheme of the integrated model and the integrated interface. The large model analyzes the requirements of the design tasks and sets calling principles based on the integrated model knowledge base and the rule base of the integrated interface, and generates corresponding prompt words according to the calling principles. S5: Construct a design task, set the genetic algorithm parameters based on the design task, and construct a multi-objective genetic algorithm; S6: Based on a multi-objective genetic algorithm, perform multi-objective optimization iterations and adaptively adjust the optimization strategy to obtain the final alloy design result.

2. The design method for refractory high-entropy alloys based on a large model according to claim 1, characterized in that, The determination of the composition space and thermodynamic content of the refractory high-entropy alloy system in S1 generates a thermodynamic dataset. Regression modeling of the thermodynamic response is performed based on this dataset to obtain a pre-trained thermodynamic surrogate model and a thermodynamic interaction interface. The thermodynamic content includes phase diagrams, the Scheil solidification process, and equilibrium phase composition, including: S11: Determine the composition space of the refractory high-entropy alloy system, wherein the composition space includes the types of elements, the content range of each element, and the calculation step size; S12: Determine the thermodynamic content to be calculated based on the application objectives. The thermodynamic content includes phase diagrams, Scheil solidification process, and equilibrium phase composition. S13: Based on the composition space and the thermodynamic content, use high-throughput thermodynamic calculation methods to perform batch calculations to obtain thermodynamic data; S14: Establish a structured specification for thermodynamic data, analyze and extract the thermodynamic data, and perform structured integration to obtain a thermodynamic dataset; S15: Regression modeling of thermodynamic response is performed based on thermodynamic dataset to obtain a pre-trained thermodynamic surrogate model. The regression modeling includes modeling using one or more machine learning models such as random forest, gradient boosting tree, support vector machine, decision tree or neural network. S16: Based on the thermodynamic data structure specifications, establish a thermodynamic interaction interface. The thermodynamic interaction interface is used to input parameters to the thermodynamic proxy model based on the large model and obtain the corresponding thermodynamic prediction results.

3. The design method for refractory high-entropy alloys based on a large model according to claim 1, characterized in that, The process S2 involves acquiring experimental data and feature sets describing the mechanical properties of refractory high-entropy alloys, training them, and obtaining a pre-trained surrogate model for predicting mechanical properties, including: S21: Obtain experimental data of refractory high-entropy alloys, wherein obtaining experimental data of refractory high-entropy alloys includes extracting and organizing data related to the target system and target mechanical property indicators from refractory high-entropy alloy related databases, published literature and proprietary experimental data; S22: Establish a structured specification for mechanical property data, and obtain a feature set for describing mechanical properties. The acquisition of the feature set for describing mechanical properties includes constructing the feature set based on materials science theory, combined with elemental characteristics, thermodynamic characteristics, phase composition characteristics, heat treatment process parameters and microstructure characteristics, and using feature engineering methods. S23: Based on the feature set and experimental data, training is performed to obtain a pre-trained mechanical performance prediction proxy model. The training includes using one or more machine learning models such as random forest, gradient boosting tree, support vector machine, decision tree or neural network, with the feature set as input and the experimental data as output. S24: Based on the structured specifications of mechanical performance data, establish a mechanical performance data interaction interface. The mechanical performance data interaction interface is used to input parameters into the thermodynamic performance prediction proxy model based on the large model and obtain the corresponding mechanical performance prediction results.

4. The design method for refractory high-entropy alloys based on a large model according to claim 1, characterized in that, The calculation model for the target physical parameters in S3 includes valence electron concentration, atomic size mismatch, mixing entropy, and density, among others. S31: Construct a calculation model for the target physical parameters. The calculation model for the target physical parameters includes constructing a calculation model for the target physical parameters based on the theory of refractory high-entropy alloys and related physical models. The target physical parameters include valence electron concentration, atomic size mismatch, mixing entropy, and density. S32: Construct a structured specification for the target physical parameter data; S33: Based on the structured specifications of the target physical parameter data, establish a target physical parameter data interaction interface. The target physical parameter data interaction interface is used to input the target physical parameter data into the calculation model of the target physical parameter based on the large model and obtain the corresponding physical parameter calculation results.

5. The design method for refractory high-entropy alloys based on a large model according to claim 1, characterized in that, The S4 component constructs an integration model and an integration interface. Based on the integration model and the integration interface, a scheduler is constructed. The scheduler includes receiving design tasks and sending them to a large model. The large model analyzes the requirements of the design tasks and determines the internal calling order and optimization scheme of the integration model and the integration interface. The large model's analysis of the design tasks includes setting calling principles based on the integration model's knowledge base and the integration interface's rule base, and generating corresponding prompts according to the calling principles, including: S41: Integrate the thermodynamic proxy model, the mechanical property prediction proxy model, and the calculation model of the target physical parameters to construct an integrated model. At the same time, construct an integrated model knowledge base based on the thermodynamic proxy model, the mechanical property prediction proxy model, and the calculation model of the target physical parameters. The integrated model knowledge base limits the model parameters and simulation space in combination with the characteristics of the refractory high-entropy alloy system. S42: Integrate the thermodynamic interaction interface, the mechanical performance data interaction interface, and the target physical parameter data interaction interface to construct an integrated interface; S43: Construct a scheduler to manage the calls to the integrated model and the integrated interface. The scheduler includes accepting design tasks and sending them to the large model. The large model analyzes the requirements of the design tasks, determines the internal calling order of the integrated model and the integrated interface, and the optimization scheme. The design tasks include the composition design range, constraints, and multi-objective tasks of the refractory high-entropy alloy to be designed. The optimization scheme includes determining the objective function, decision variable representation, constraints, and genetic operators of the multi-objective optimization problem. The large model analyzes the requirements of the design tasks, including setting calling principles based on the integrated model knowledge base and the rule base of the integrated interface, and generating corresponding prompts according to the calling principles.

6. The design method for refractory high-entropy alloys based on a large model according to claim 1, characterized in that, The construction design task of S5 involves setting genetic algorithm parameters based on the design task and constructing a multi-objective genetic algorithm, including: S51: Based on the composition design range, constraints and multi-objective tasks of the refractory high-entropy alloy to be designed, a design task is constructed. The multi-objective tasks include quantifiable performance indicators of the refractory high-entropy alloy to be designed under target service conditions. The performance indicators include at least one or more of thermal stability-related indicators, strength and hardness-related mechanical property indicators and density-related physical parameters. The design task defines the composition design space. S52: Set the genetic algorithm parameters and construct a multi-objective genetic algorithm for searching the Pareto front of the target performance.

7. The design method for refractory high-entropy alloys based on a large model according to claim 1, characterized in that, The S6 algorithm, based on a multi-objective genetic algorithm, performs multi-objective optimization iterations and adaptively adjusts the optimization strategy to obtain the final alloy design results, including: S61: Based on the design task, the multi-objective genetic algorithm is invoked to randomly initialize N individuals within the given composition design space, where each individual represents a candidate alloy composition scheme; S62: Input each individual into the ensemble model through the scheduler to obtain the multidimensional performance index of each individual; S63: Evaluate the merits and demerits of each individual based on the aforementioned multidimensional performance indicators, and determine the direction of evolution; S64: Select parent individuals from the current population, perform crossover and mutation operations to generate N new offspring individuals; S65: Through multiple generations of iterative evolution, the population is continuously updated until a Pareto front solution set that approximates the target performance is obtained; S66: Analyze the Pareto front solution set obtained based on the large model. If the design goal or constraints are not met, adaptively adjust the design task and genetic algorithm parameters until the preset convergence condition is met or the predetermined number of iterations is reached to obtain the optimal design task. S67: Combine the optimal design task with the corresponding multi-objective performance prediction value and the difference or deviation index from the target performance to obtain the final alloy design result.

8. A design apparatus for refractory high-entropy alloys based on a large model, wherein the apparatus is used to implement the design method for refractory high-entropy alloys based on a large model as described in any one of claims 1-7, characterized in that, The device includes: Thermodynamics module: used to determine the composition space and thermodynamic content of refractory high-entropy alloy systems, generate thermodynamic datasets, perform regression modeling on thermodynamic responses based on thermodynamic datasets, and obtain pre-trained thermodynamic surrogate models and thermodynamic interaction interfaces. The thermodynamic content includes phase diagrams, Scheil solidification process, and equilibrium phase composition. Mechanical properties module: used to acquire experimental data and feature sets describing the mechanical properties of refractory high-entropy alloys, train them, and obtain a pre-trained mechanical property prediction surrogate model; Physical Parameters Module: Used to construct computational models for target physical parameters, including valence electron concentration, atomic size mismatch, mixing entropy, and density; Integrated Model Module: Used to build an integrated model, build an integrated interface, and build a scheduler based on the integrated model and the integrated interface. The scheduler construction includes accepting design tasks and sending them to the large model. The large model analyzes the requirements of the design task and determines the internal calling order and optimization scheme of the integrated model and the integrated interface. The large model analyzes the requirements of the design task and sets calling principles based on the integrated model knowledge base and the rule base of the integrated interface, and generates corresponding prompt words according to the calling principles. Multi-objective genetic algorithm module: used to construct design tasks, set genetic algorithm parameters based on design tasks, and construct multi-objective genetic algorithms; Optimization module: Used to perform multi-objective optimization iterations based on multi-objective genetic algorithms and adaptively adjust the optimization strategy to obtain the final alloy design result.

9. A design device for refractory high-entropy alloys based on a large model, characterized in that, The processor for designing refractory high-entropy alloys based on a large model; a memory storing computer-readable instructions that, when executed by the processor, implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium contains program code that can be invoked by a processor to execute the method as described in any one of claims 1 to 7.

Images