基于改进粒子群算法神经网络的SS304不锈钢激光拼焊焊缝成形预测  

Prediction of weld seam formation of SS304 stainless steel laser tailor welding based on improved particle swarm algorithm neural network

作　　者：张群莉 靳威威 屠夏琦 陈智君 薛华军[2,3] 吴许祥 倪振兴 王成[2,3] 郭星 ZHANG Qunli;JIN Weiwei;TU Xiaqi;CHEN Zhijun;XUE Huajun;WU Xuxiang;NI Zhenxing;WANG Cheng(Laser Advanced Manufacturing Research Institute,Zhejiang University of Technology,Hangzhou 310023,Zhejiang China;Jiangsu Yawei Machine Tool Co.,Ltd.,Yangzhou 225200,Jiangsu China;Jiangsu Province Key Laboratory of Sheet Metal Intelligent Equipment,Yangzhou 225200,Jiangsu China)

机构地区：[1]浙江工业大学激光先进制造研究院,浙江杭州310023 [2]江苏亚威机床股份有限公司,江苏扬州225200 [3]江苏省金属板材智能装备重点实验室,江苏扬州225200 [4]不详

出　　处：《锻压装备与制造技术》2021年第2期106-110,共5页China Metalforming Equipment & Manufacturing Technology

摘　　要：为了研究SS304不锈钢板拼焊的焊缝成形效果,搭建了激光焊接试验平台和数据采集系统,并通过试验采集的数据样本作为神经网络的训练样本。建立了基于改进粒子群算法神经网络的激光拼焊焊缝成形预测模型,利用遗传算法群体搜索策略改进粒子群算法获得神经网络修正后的初始权值及阈值,以激光拼焊工艺参数激光功率、扫描速度、板厚和离焦量作为神经网络的输入,焊缝高度、焊缝宽度、抗拉强度作为输出;结果表明遗传算法改进的粒子群算法神经网络能够对焊缝高度、焊缝宽度、抗拉强度进行准确的预测。In order to study the weld seam forming effect of tailor-welded SS304 stainless steel plate,a laser welding test platform and a data acquisition system have been built and the data samples collected through the test have been used as the training samples of the neural network.The laser tailor welding seam formation prediction model based on improved particle swarm algorithm neural network has been established.The genetic algorithm group search strategy has been used to improve the particle swarm algorithm to obtain the initial weights and thresholds after the neural network correction.The laser tailor welding process parameters,laser power,scanning speed,plate thickness and defocus amount have been used as the input of the neural network while the height,width and tensile strength of the weld seam have been used as the output.The results show that the particle swarm algorithm neural network improved by genetic algorithm can make accurate prediction of the height,width and tensile strength of weld seam.

关 键 词：焊缝 遗传算法 粒子群算法 神经网络 预测 

分 类 号：TG456.7[金属学及工艺—焊接] TP183[自动化与计算机技术—控制理论与控制工程]