基于Dual-Kriging模型多目标稳健设计在板料拉深成形中的应用  

The application of multi-objective robust design based on Dual-Kriging in the sheet drawing process

作　　者：胡静[1] 谢延敏[1] 王智[1] 王新宝[1] 

机构地区：[1]西南交通大学机械工程学院,成都610031

出　　处：《塑性工程学报》2013年第3期37-42,共6页Journal of Plasticity Engineering

基　　金：国家自然科学基金资助项目(51005193)

摘　　要：板料拉深成形工艺中,由于设计参数波动的影响,导致在传统最优工艺条件下成形件质量不够稳健。该文采用板料成形有限元仿真、代理模型,并结合稳健设计思想,提出基于Dual-Kriging模型的稳健设计方法。筛选影响成形质量的主要模具几何、工艺参数,并确定其相应波动范围;基于有限元仿真,运用Kriging模型,建立设计参数与成形质量的函数关系;根据参数波动情况,基于第一重Kriging模型以及Monte Carlo模拟,运用Kriging模型,建立设计参数与成形质量期望与方差的第二重模型;运用基于拥挤距离的粒子群算法,以及灰色关联分析进行多目标优化,完成设计参数的稳健设计。将该方法运用于NUMISHEET 93的方盒件拉深成形,与传统确定性优化设计相比,可有效减少成形质量的波动范围以及不确定性分析的计算量,提高成形件的稳健性。In the sheet drawing process, the forming parts are not enough robust under the traditional optimal process condition, due to the fluctuation of the design parameters. Therefore, combining the finite element simulation of the sheet metal forming with the ideas of the robust design, the robust design based on Dual-Kriging model is proposed. Choose the main geometry, process parameters which has great effect on the forming quality and determine the corresponding fluctuation range; establish the functions between the design parameters and the forming quality using the Kriging model based on the finite element simulation; the second layer Kriging model of forming quality expectation and variance are built based on the first layer Kriging model and Monte Carlo simulation according to the parameter fluctuations; use the particle swarm optimization based on the crowded dis- tance and gray relational analysis to optimize and complete the process of the robust design of the parameters. This method is ap- plied into the square cup drawing in NUMISHEET 93 and compared to the traditional deterministic optimization. The forming quality fluctuation and the calculation of the uncertainly analysis are greatly decreased and the robustness of the forming part is im- proved.

关 键 词：板料成形 Dual—Kriging模型 多目标稳健优化 粒子群算法 

分 类 号：TG386.1[金属学及工艺—金属压力加工]