Method and device for designing refractory high-entropy alloy based on large model

By integrating thermodynamic, mechanical property, and physical parameter models into a collaborative design approach, a large model is used for multi-objective optimization of refractory high-entropy alloys. This solves the problems of model fragmentation and reliance on manual intervention, and achieves efficient multi-objective performance balance.

CN122135833APending Publication Date: 2026-06-02UNIV OF SCI & TECH BEIJING
0 Cites 0 Cited by

Patent Information

Application Number
CN202610041329.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-01-13
Publication Date
2026-06-02

AI Technical Summary

Technical Problem

Existing design methods for refractory high-entropy alloys suffer from problems such as model fragmentation, low efficiency of multi-objective optimization, and high dependence on human intervention in the optimization process. They lack a unified model integration and collaboration mechanism, making it difficult to efficiently obtain reasonable Pareto compromise solutions under high-dimensional composition space and complex constraints.

Method used

A design method for refractory high-entropy alloys based on a large model is constructed. By integrating a thermodynamic surrogate model, a mechanical property prediction surrogate model, and a physical parameter calculation module, a multi-objective genetic algorithm is used for optimization to achieve synergistic optimization of composition design and multi-objective performance.

Benefits of technology

It improves the efficiency of multi-objective optimization, simplifies model calling and search steps, and the obtained candidate alloys perform better in terms of multi-objective performance balance, reducing the dependence on manual work.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a refractory high-entropy alloy design method and device based on a large model, and relates to the technical field of alloy material design. The method comprises the following steps of: firstly, respectively constructing pre-trained agent models of thermodynamics (covering a phase diagram, a solidification process and the like) and mechanical properties by generating and collecting data, and establishing a calculation model of physical parameters such as valence electron concentration, density and the like; then, the models are integrated into an integrated model, and a large language model is introduced as an intelligent scheduling core. The large model is responsible for analyzing a design task and intelligently planning and calling each model interface according to a knowledge base and a rule base. And finally, based on a set design target, a multi-target genetic algorithm is started to carry out iterative optimization and adaptive adjustment, and a final alloy design scheme meeting multiple performance requirements is efficiently output. According to the method, intelligent multi-target design of the refractory high-entropy alloy can be realized, and the problem of high dependency on manpower and the like is solved.
Need to check novelty before this filing date? Find Prior Art