摘要
采用6000W光纤激光器对1mm、2mm、4mm、6mm厚的SS304不锈钢进行激光焊接工作,同时结合BP神经网络与数据库技术,设计了一种基于BP神经网络的激光焊接工艺参数优化数据库系统,提高激光焊接工艺的精度和质量。BP网络训练次数达到5000次以上时,系统输出精度优于0.015mm,同时,焊接熔宽比与焊缝宽度的实测值与网络输出值最大误差分别为0.03mm与0.07mm,焊缝宏观形貌美观、微观组织金相优良,实现了高精度、高质量、高稳定度的设计目标。
A 6000 W fiber laser generator has been used for laser welding of 1 mm, 2 mm, 4 mm, 6 mm thick SS304 stainless steel, in combination with BP neural network and database technology, a laser welding process parameter optimization database system based on BP neural network has been designed to improve the precision and quality of the laser welding process. When the number of BP network training reaches more than 5000 times, the output accuracy of the system is better than 0.015 mm. At the same time, the maximum error between the actual measured value of the welding melt width ratio and the weld width and the network output value are respectively 0.03 mm and 0.07 mm. The welding seam has the beautiful macroscopic appearance and excellent microstructure and metallographic structure. The design goals of high precision, high quality and high stability have been achieved.
作者
吴许祥
王成
薛华军
沈店祥
WU Xuxiang;WANG Cheng;XUE Huajun;SHEN Dianxiang(Jiangsu Yawei Machine Tool Co.,Ltd.,Yangzhou 225200,Jiangsu China;Jiangsu Province Key Laboratory of Sheet Metal Intelligent Equipment,Yangzhou 225200,Jiangsu China)
出处
《锻压装备与制造技术》
2021年第6期77-82,共6页
China Metalforming Equipment & Manufacturing Technology
