摘要
为了提高激光焊接工艺的精度和质量,采用BP神经网络与数据库相结合的技术,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,并进行了实验。结果表明,BP网络训练次数达到3000次以上时,系统输出精度优于0.015mm,同时,焊接熔深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0.007mm,实现了高精度、高质量、高稳定度的设计目标,具有很好的应用前景。
In order to improve the accuracy and quality of laser welding process,using the technology of combination of BP neural network and database,an optimized database system of technologic parameters of laser welding process based on BP neural network is designed,and an experiment is per formed.The results indicate that the accuracy of the system output is superior to 0.015 mm,if the training number of BP network more than 3000,and the errores of the measured value and the network output on weld depth and weld width are 0.03 mm and 0.007 mm,achieves the design goal of high precision,high quality,high stability,and has certain application prospects.
出处
《激光杂志》
北大核心
2015年第4期182-184,共3页
Laser Journal
基金
山西省软科学研究项目资助课题(2012041010-03)
关键词
激光
BP
数据库
优化
Lasers
BP
Database
Optimization
