基于有限元法的梯度材料零件多目标优化设计  

Multi-Objective Optimal Design of Graded Material Part Based on Finite Element Method

作　　者：赵韩[1] 吴振华[1] 张冰战[1] 

机构地区：[1]合肥工业大学机械与汽车工程学院,合肥230009

出　　处：《系统仿真学报》2014年第1期169-172,共4页Journal of System Simulation

基　　金：国家"863"重大专项资助课题(2008AA11A139)

摘　　要：针对由两相材料在空间中连续过渡组成的梯度材料零件的多目标优化设计,建立了基于有限元法的梯度材料分析模型,构造了基于网格节点的设计变量,以材料体积分数进行实值编码,采用了外部存档式微遗传算法AMGA进行优化计算,进而提出了梯度材料多目标进化方法进行零件的设计。仿真实例的计算结果表明,该设计方法能够得到满足设计目标和约束条件的解,求解效率和精度较高。Aiming at multi-objective optimal design of graded material part composed of two-phase materials transmitting smoothly in space, the analysis model for graded material was established based on the finite element method. Real-valued volume fractions of material were encoded according to design variables constructed on the mesh nodes. AMGA （Archive-based Micro Genetic Algorithm） was used to perform the optimization and a method of multi-objective graded material evolution was proposed to design the corresponding parts. An example for simulation and the computation results show that solutions can be found efficiently and accurately while satisfying the design objectives and constraints.

关 键 词：梯度材料 有限元法 多目标优化 实值编码 外部存档式微遗传算法 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]