文摘
|
基于人工神经网络技术对覆盖件模具表面激光硬化虚拟过程的仿真建模,结合几何因素分析了模型的主要影响参数,对BP网络的结构和训练进行了说明。预测了激光表面硬化的加工效果 (表面硬度、硬化层深、相对耐磨性和表面粗糙度),实现激光加工工艺参数的优化,为实际生产和加工提供了依据。并以C语言为开发语言,利于实现各平台间的集成。 |
其他语种文摘
|
The artificial neural network methodology has been developed for the virtual simulation of the laser surface hardening process for blanket dies. Laser-control parameters and surface geometrical configurations are taken as the main input data. BP network model is built after the training of sample data to predict the hardening effect including hardness, hardening depth, relative wear resistance and surface roughness. Comparison of predicted and experimental data has confirmed the accuracy of the neural network approach to optimize the technological parameters and provide effective foundation in practical production. The program of the neural network is written in C language due to its versatility. |
来源
|
金属热处理
,2005,30(3):61-64 【核心库】
|
关键词
|
人工神经网络
;
覆盖件模具
;
激光表面硬化
;
过程仿真
|
地址
|
中国科学院力学研究所, 北京, 100080
|
语种
|
中文 |
文献类型
|
研究性论文 |
ISSN
|
0254-6051 |
学科
|
金属学与金属工艺 |
基金
|
中国科学院大型仪器设备研制项目
;
中国科学院知识创新工程重大项目
|
文献收藏号
|
CSCD:2019729
|