Mechanical properties of SMA W （shielded metal arc welding） weld metal （ yield strength higher than 900 MPa ） with systemazic additions of copper （ up to 1.48 wt% ） were tested, The microstructure and precipitat...Mechanical properties of SMA W （shielded metal arc welding） weld metal （ yield strength higher than 900 MPa ） with systemazic additions of copper （ up to 1.48 wt% ） were tested, The microstructure and precipitates in different regions were analyzed by optical microscope and transmission electron microscope, The results indicate that copper improves the low temperature toughness of weld metal when the copper content is low and reaches the peak value 48 J （ at - 50℃ ） with 0. 2 wt% copper additions. When the content is high the copper precipitates as 8-Cu phase in the reheat zone of middle beads. These precipitates improve the strength of the weld metal evidently （ yield strength up to 975 MPa） without obvious effect on the low temperature toughness. The copper within 1.1 wt% content can improve the strength without toughness loss.展开更多
重大装备制造中厚板机器人多层多道焊(multi-layer and multi-pass welding,MLMPW)一直是热点和难点,而实现机器人MLMPW的核心是对其熔池的获取、监控并分类.为了提高MLMPW的自动化和智能化,有必要开发一个熔池图像在线分类系统.针对焊...重大装备制造中厚板机器人多层多道焊(multi-layer and multi-pass welding,MLMPW)一直是热点和难点,而实现机器人MLMPW的核心是对其熔池的获取、监控并分类.为了提高MLMPW的自动化和智能化,有必要开发一个熔池图像在线分类系统.针对焊接过程中的熔池图像提出了一种新的MLMPW熔池分类方法——基于视觉注意的(SENet)VGGNet熔池分类方法.为了提高效率和精度,引入迁移学习中的预训练模型到网络训练过程中.因为针对中厚板多层多道熔池研究较少,导致熔池公开数据集较少,为了应对这一问题,需要对数据集进行增广.结果表明,提出的模型可快速有效的对七类MLMPW熔池进行准确分类,预测精度可达到98.39%.展开更多
文摘Mechanical properties of SMA W （shielded metal arc welding） weld metal （ yield strength higher than 900 MPa ） with systemazic additions of copper （ up to 1.48 wt% ） were tested, The microstructure and precipitates in different regions were analyzed by optical microscope and transmission electron microscope, The results indicate that copper improves the low temperature toughness of weld metal when the copper content is low and reaches the peak value 48 J （ at - 50℃ ） with 0. 2 wt% copper additions. When the content is high the copper precipitates as 8-Cu phase in the reheat zone of middle beads. These precipitates improve the strength of the weld metal evidently （ yield strength up to 975 MPa） without obvious effect on the low temperature toughness. The copper within 1.1 wt% content can improve the strength without toughness loss.
文摘重大装备制造中厚板机器人多层多道焊(multi-layer and multi-pass welding,MLMPW)一直是热点和难点,而实现机器人MLMPW的核心是对其熔池的获取、监控并分类.为了提高MLMPW的自动化和智能化,有必要开发一个熔池图像在线分类系统.针对焊接过程中的熔池图像提出了一种新的MLMPW熔池分类方法——基于视觉注意的(SENet)VGGNet熔池分类方法.为了提高效率和精度,引入迁移学习中的预训练模型到网络训练过程中.因为针对中厚板多层多道熔池研究较少,导致熔池公开数据集较少,为了应对这一问题,需要对数据集进行增广.结果表明,提出的模型可快速有效的对七类MLMPW熔池进行准确分类,预测精度可达到98.39%.
