摘要
针对板料拉深成形过程中摩擦系数较难确定的客观情况 ,提出一种基于人工神经网络的能够快速识别摩擦系数的方法 .在概述解析法确定摩擦系数的基础上 ,建立了识别摩擦系数的人工神经网络模型 ,用拉深试验机测取的摩擦系数、压边力、拉深力和凸模行程组成样本对其训练 ,实现了对摩擦系数的快速识别 .从而就可以根据摩擦系数的波动 ,适时调整控制参数 ,以最佳的工艺参数来完成板料的整个拉深成形过程 .实验结果表明 。
A new method based on artificial neural network to identify the friction coefficient quickly was proposed according to the state of the friction coefficient in sheet metal deep drawing process. Artificial neural network model is established and trained by specimens composed of friction coefficient, blank holder force, drawing force and punch displacement. The friction coefficient is identified quickly on the basis of its analytic description. So the control parameter can be adjusted constantly according to the friction coefficient fluctuation to finish the whole process of sheet metal deep drawing with optimum technological parameters. Experimental result shows the friction coefficient in sheet metal deep drawing can be identified quickly and precisely using artificial neural network. It is an effective method to determine the friction coefficient in sheet metal deep drawing.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2001年第7期969-971,976,共4页
Journal of Shanghai Jiaotong University
