基于近似替代模型的多工位高速锻造预锻模具优化  

Optimization of preform of high-speed multi-stage forging based on surrogate model methodology

作　　者：吴彦骏[1] 赵震[1] 梁艳迁[1] 胡成亮[1] 高崇晖[1] 

机构地区：[1]上海交通大学模具CAD国家工程研究中心,上海200030

出　　处：《塑性工程学报》2009年第4期79-84,共6页Journal of Plasticity Engineering

摘　　要：多工位高速锻造是一种节能减耗的先进成形技术,与传统的单工位多工序锻造有较大区别,不合适的预锻工位模具设计在生产中会造成载荷过大、锻件充填质量差、折叠等问题。文章提出了基于近似替代模型的多工位高速锻造预锻模具优化设计方法,采用拉丁超立方抽样方法构造初始样本点集,应用有限元软件获得响应值,并应用Kriging模型和BP神经网络模型建立真实问题的替代模型,来近似拟合预锻工位模具形状与成形载荷、成形质量之间的关系,并结合基于惩罚函数法的遗传算法求得最优解。以3工位高速锻造问题为例,经过优化取得了良好的效果,与初始设计相比,成形质量显著提高,成形载荷下降了40%左右。The high-speed multi-stage forging is an advance forming technique. It is different from the traditional die forging process. The improper preform design will lead to some problem in actual production, such as too large toad, underfill, folding and so on. In order to obtain the optimal design, an new optimization method based on surrogate models is researched in this paper. This methodology applies the Latin Hypercube Sampling （LHS） to generate initial experiment samples. Then the values of response functions are obtained by numerical simulation. Afterwards, the kriging model and BP neural network are established the surrogate models to fit the relationships between variables and responses. These surrogate models are used in optimization calculation to obtain the optimal design by using Genetic Algorithm based on penalty function. A 3-station high-speed forging case was studied with the above-mentioned method. After optimization, the quality of the product was improved significantly and the forming load decreased more than 400% compared to the initial design.

关 键 词：多工位高速锻造 拉丁超立方采样 KRIGING BP神经网络 优化 

分 类 号：TG316.3[金属学及工艺—金属压力加工]