二次方程
数学优化
材料科学
微晶
航程(航空)
二次规划
马氏体
转化(遗传学)
形状记忆合金
计算机科学
应用数学
序列二次规划
相(物质)
能量(信号处理)
高效能源利用
钥匙(锁)
钛镍合金
背景(考古学)
计算机模拟
半定规划
最优化问题
应变能
结构工程
能量转换
算法
数值分析
数学
标识
DOI:10.1088/1361-665x/ae5e32
摘要
Abstract We present efficient numerical strategies for the implementation of a recently proposed three-dimensional, micromechanically consistent energy model for polycrystalline shape-memory alloys. The model accounts for intra- and inter-granular energy constraints governing martensitic variant selection and the formation of austenite–martensite microstructures. Its practical use has so far been limited by the high computational cost arising from nested, high-dimensional, and nonconvex optimization problems that must be solved to evaluate the effective energy. To overcome this difficulty, we introduce systematic approximations and algorithmic techniques that significantly reduce computational cost while preserving accuracy. The proposed approximations are formulated in terms of semidefinite programming and quadratic programming problems, which can be solved using well-established numerical solvers. Their accuracy and efficiency are assessed on a range of test cases involving different numbers of active martensitic variants. The proposed approach is integrated into a full polycrystalline simulation framework and applied to the prediction of the superelastic response of a three-dimensional NiTi polycrystal under uniaxial tension. The simulation results reproduce key experimental features on phase transformation and strain localization.
科研通智能强力驱动
Strongly Powered by AbleSci AI