三辊行星轧制工艺参数优化技术研究  

Specification optimization for three-roller planetary rolling

作　　者：刘劲松[1] 蔚敬泽[1] 岳峰丽[1] 田春霞[1] 

机构地区：[1]沈阳理工大学材料科学与工程学院,辽宁沈阳110168

出　　处：《中国工程机械学报》2008年第3期340-343,共4页Chinese Journal of Construction Machinery

基　　金：辽宁省博士启动基金资助项目(20071026)

摘　　要：三辊行星轧制是铜合金管材加工的主要工序之一,对其工艺参数进行优化具有重要意义.采用将正交试验设计试验方案、BP神经网络建立映射关系和遗传算法寻优相结合的优化技术,可以获得三辊行星轧制最佳工艺参数组合.这种技术既利用了正交试验的多因素试验特性,神经网络的自学习、自适应非线性映射与预测功能,又利用了遗传算法的自然选择和进化的全局寻优特性,可以为实际生产提供可靠的参数依据.Due that the three-roller planetary rolling is a critical stage to produce copper alloy pipes, it is significant to optimize related specifications. In this regard, a method, which integrates orthogonal experiment design, BP neural networks and genetic algorithms, is developed for the optioning, mapping and optimization purposes. Afterwards, the optimal combination of specifications can be obtained. Therefore, this approach has revealed its reliability in practice by taking such advantages of multi-criteria testing via orthogonal experiment design, self-learning and adaptive nonlinear mapping and forecasting via neural networks, and natural selection and evolution via genetic algorithms.

关 键 词：三辊行星轧制 正交实验设计 神经网络 遗传算法 

分 类 号：TG386[金属学及工艺—金属压力加工]