基于数据驱动的颗粒增强金属基复合材料的本构建模方法及应力预测方法

本发明属于数据驱动计算力学和材料本构模型技术领域，公开了一种基于数据驱动的颗粒增强金属基复合材料的本构建模方法及应力预测方法，该方法包括：根据颗粒增强金属基复合材料的结构建立代表性体积单元模型；输入多种不同组分的颗粒增强金属基复合材料的杨氏模量、泊松比、屈服应力、塑性应变，并对其施加不同路径下的载荷，根据有限元计算结果提取划分数据集；将数据集输入到BP神经网络中完成线下训练过程；并根据训练结果实现基于神经网络的应力预测和一致切线模量的更新，获得颗粒增强金属基复合材料本构模型。本发明可以快速准确地建立颗粒增强金属基复合材料的本构模型，更加准确地描述其力学响应行为，提高数值模拟的精度。

This invention falls within the technical field of data-driven computational mechanics and material constitutive modeling, and discloses a data-driven constitutive modeling method and stress prediction method for particle-reinforced metal matrix composites (PRMMCs). The method comprises the following steps: firstly, establishing a representative volume element (RVE) model based on the structure of the particle-reinforced metal matrix composite; secondly, inputting the Young's modulus, Poisson's ratio, yield stress and plastic strain of particle-reinforced metal matrix composites with a range of compositions, applying loads under different loading paths to them, and extracting and partitioning the dataset based on finite element calculation results; thirdly, inputting the dataset into a backpropagation (BP) neural network to complete the offline training process; finally, realizing neural network-based stress prediction and consistent tangent modulus update according to the training results, so as to obtain the constitutive model of particle-reinforced metal matrix composites. This invention can quickly and accurately establish the constitutive model of particle-reinforced metal matrix composites, more accurately describe their mechanical response behavior, and improve the accuracy of numerical simulation.

搜集汇总

背景与挑战

背景概述

该数据集聚焦于颗粒增强金属基复合材料的本构建模和应力预测，采用数据驱动方法结合BP神经网络进行训练，旨在快速建立准确的本构模型以提升数值模拟精度。其特点包括利用代表性体积单元模型和有限元计算生成数据，有效描述材料的力学响应行为。