摘要
以改进型遗传神经网络为理论基础,建立以焊接电流、焊接电压、焊接速度和焊后冷却速度为输入,以横向收缩变形和角变形为输出的手工电弧焊(SMAW)焊接变形预测模型,并通过正交试验法设计试验,验证了该焊接变形预测模型。结果表明:与基于传统遗传神经网络的SMAW焊接变形预测模型相比,本模型预测精度更高,速度更快,具有更高的工程应用价值。
Based on improved genetic neural network, creating a welding deformation prediction model of SMAW with the inputs of welding current, welding voltage, welding speed and cooling rate and the outputs of transverse shrinkage deformation and angular distortion, an experiment was designed by orthogonal experimental. The prediction model was tested .The results of this experiment shows that： compared with traditional prediction model of welding deformation, the model gets higher accuracy, speed and engineering value.
出处
《热加工工艺》
CSCD
北大核心
2015年第1期208-210,共3页
Hot Working Technology
关键词
遗传神经网络
电弧焊
焊接变形
genetic neural network
electric arc welding
welding deformation
