焊道形貌特征的BP神经网络建模与预测  

Welded Surface Morphology Modeling and Prediction Based on BP Neural Network

作　　者：黄无云 刘益剑[1] 刘宗熙 杨继全[1] 朱钊伟 谢非[1] 史建军 Huang Wuyun;Liu Yijian;Liu Zongxi;Yang Jiquan;Zhu ZhaoWei;Xie Fei;Shi Jianjun(NARI School of Electrical and Automation Engineering,Nanjing Normal University,Nanjing 210023,China;Nanjing Zhongke Yuchen Laser Technology Co.,Ltd.,Nanjing 210038,China)

机构地区：[1]南京师范大学南瑞电气与自动化学院,江苏南京210023 [2]南京中科煜宸激光技术有限公司,江苏南京210038

出　　处：《南京师范大学学报（工程技术版）》2021年第1期1-7,共7页Journal of Nanjing Normal University(Engineering and Technology Edition)

基　　金：国家重点研发计划项目(2017YFB11032002);江苏省科技成果转化专项资金项目(BA2020004);2020年江苏省省级工业和信息产业转型升级专项资金项目(智能化金属增减材制造装备).

摘　　要：良好的焊道是成功进行电弧增材制造的保障,其受到焊接电流、电压、扫描速度、送丝速度等多种参数影响.提出了以焊道高度、宽度为形貌特征的4输入2输出BP神经网络模型,并利用PSO进行了神经网络权值的优化求解.实验结果表明,设计的BP神经网络实现了对焊道形貌的预测,为后续电弧增材制造的实时预测与控制提供了模型基础.A good weld bead provides a guarantee for successful wire arc additive manufacturing,which is affected by various parameters such as welding current,voltage,scanning speed,and wire feed speed.A four-input and two-output BP neural network model is proposed,which is applied to the surface morphology characteristic identification of the weld bead height and width.The particle swarm optimization(PSO)algorithm is used to optimize the neural network weight.Experimental results show that the BP neural network designed in this paper realizes the prediction of the weld bead morphology and that it provides a model basis for the real-time prediction and control of subsequent arc additive manufacturing.

关 键 词：焊道形貌 BP神经网络 优化求解 粒子群算法 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]