基于RBF人工神经网络的机器人焊缝成形质量预测  

Quality prediction of robot welding seam molding basic on RBF artificial neural network

作　　者：乌日根[1] 王玉荣[1] WURigen;WANG Yurong(Baotou Vocational Technical College,Baotou 014030,Inner Mongolia Autonomous Region,China)

机构地区：[1]包头职业技术学院材料工程系,内蒙古包头014030

出　　处：《焊接技术》2022年第9期102-105,I0008,共5页Welding Technology

基　　金：内蒙古自治区高等学校科学研究重点项目(NJZZ21064)。

摘　　要：采用正交试验方法获取典型样本数据,通过Matlab软件建立焊缝成形质量与主要焊接工艺参数之间的人工神经网络模型,预测研究焊缝填充层焊接工艺参数对焊缝成形质量的影响规律。结果表明,RBF神经网络能够较为精确反映多层焊填充层机器人焊接工艺参数与焊缝余高及宽度之间的非线性映射关系,可用于机器人多层焊焊接工艺参数优化和焊缝成形质量的控制;随着填充层焊接工艺参数值提高,焊缝余高降低,焊缝宽度增加,当继续增大焊接工艺参数时,焊接电流和电弧电压的影响相对更加显著,而焊接速度的影响却减小。The paper predicted the influence of welding parameters of packed layer on the quality of appearance of weld of based on the typical sample data obtained from orthogonal test and the artificial neural network model between the quality and the parameters built by MATLAB.The result showed that the RBF neural network could reflect the non-linear mapping relationships more accurately among robot welding parameters,the weld reinforcement and the width,which could be used in the control of robot multilayer welding parameters and the welding seam quality of appearance of weld.With the increase of the parameter values,the weld reinforcement decreased and the width increased.When the values continued to improve,the effect of welding current and arc voltage was relatively more significant,while the effect of welding speed was weakened.

关 键 词：正交试验 RBF神经网络 焊接工艺参数 焊缝成形 

分 类 号：TG441.4[金属学及工艺—焊接] TG183