摘要
针对金属板料抛喷丸加工成形与变形控制问题,研究了基于等效弯矩载荷建立喷丸数值模拟的方法。为实现基于喷丸工艺参数预测等效载荷,提出基于神经网络训练与学习的方法。以试验数据为样本数据,以模拟数据动态修正的方法,动态获取工艺参数与等效载荷的关系模型。使用VC/OpenGL编程,基于AFT方法实现了有限元自动网格剖分,并针对多抛头交错布置的特点,采用基于时间段按网格细分与迭加原理实现了板料的动态模拟,得到板料的最终变形与应力云图。
In order to solve the problem of forming and deformation control in the process of shot blasting and shot peening of metal sheet,a method of numerical simulation based on equivalent moment is presented.For forecasting the equivalent moment by process parameters,we propose the method of training and study based on neural networks.Using test data as the sample data,we build the relation model of process parameters and equivalent moment by modifying the simulation data dynamically.We use VC/OpenGL for automatic mesh generation by advancing front technique （ AFT） .Because of the staggered layout of multi-throwing,gridding subdivision with the passage of time and iteration are used to get the final deformation and nephogram of stress.
出处
《机械科学与技术》
CSCD
北大核心
2010年第6期836-840,共5页
Mechanical Science and Technology for Aerospace Engineering
关键词
抛喷丸加工
等效弯矩
数值模拟
神经网络
动态获取
shot blasting and shot peening
equivalent moment
numerical simulation
neural networks
