Transactions of Nonferrous Metals Society of China The Chinese Journal of Nonferrous Metals

您目前所在的位置:首页 - 期刊简介 - 详细页面

中国有色金属学报

ZHONGGUO YOUSEJINSHU XUEBAO

第24卷    第3期    总第180期    2014年3月

[PDF全文下载]        

    

文章编号:1004-0609(2014)03-0678-11
镀锌钢/6016铝合金激光焊的BP神经网络工艺优化及 组织和性能
周惦武,乔小杰,张丽娟,李 升

(湖南大学 汽车车身先进设计制造国家重点实验室,长沙 410082)

摘 要: 为了抑制或减少镀锌钢/6016铝合金在激光焊过程中生成Fe/Al脆性金属间化合物,采用光纤激光器,不添加任何钎料,对汽车车身用镀锌钢与6016铝合金平板试件进行激光搭接焊试验。利用ANSYS有限元分析软件,建立钢/铝激光焊热源模型。基于模拟计算与温度同步测量,获取不同工艺条件下焊接熔深,将其作为训练样本,利用BP神经网络,建立激光焊条件下焊接功率、焊接速度、离焦量与焊接熔深之间的非线性映射关系,优化焊接工艺参数。采用卧式金相显微镜、扫描电镜、X射线衍射仪、微机控制电子万能试验机等手段对优化工艺条件下焊接接头各区域的金相组织、断口形貌、界面元素分布、主要物相和接头力学性能进行研究。结果表明:模拟计算获得工件表面距焊缝中心不同位置温度与实验同步测量温度基本吻合,所建热源模型能反映激光焊实际过程特点;通过神经网络方法建立模型预测不同工艺参数下焊接熔池深度的相对误差控制在10%以内;优化工艺条件下,焊缝横截面铝熔化,钢少量熔化,液态铝在钢表面润湿铺展良好,钢/铝界面层形成厚度约为9 μm、由FeAl和Fe3Al组成的金属间化合物层,焊接接头平均抗剪强度为27.70 MPa,断裂形貌表现为准解理和韧性的混合型断裂特征。

 

关键字: 6016铝合金;镀锌钢;激光焊;模拟计算;BP神经网络;组织;性能

Parameters optimization of laser welding process of galvanized steel and 6016 aluminum alloy based on BP neural network and its microstructure and mechanical properties
ZHOU Dian-wu, QIAO Xiao-jie, ZHANG Li-juan, LI Sheng

State Key Laboratory of Advanced Design and Manufacturing for Vehicle Body,
Hunan University, Changsha 410082, China

Abstract:The laser lap welding test on the galvanized steel and the 6016 aluminum alloy with tie fiber laser and no brazing filler metal addition was carried out in order to inhibit or reduce the formation of Fe/Al brittle intermetallic compound. The laser welding heat source model of steel and aluminum was established by using ANSYS software. The penetration depths under different process parameters were given based on the simulation and experiment results, the nonlinear relationship between the process parameters, such as laser power, welding velocity and focal position and the penetration depth was established by the BP network, the welding parameters were also optimized. The microstructure, fracture morphology, element distribution, the main phase of the welded joint regional and mechanical properties of welded joints were studied by optical microscopy, scanning electron microscopy, X-ray diffractometry and microcomputer tensile testing machine. The results show that the simulation temperatures are consistent with the measured temperatures of different positions on the workpiece surface and the heat source model can be applied to the actual laser welding process, the relative error of the penetration depth under different process parameters can be controlled within 10% according to the trained network forecast. Under the optimal process conditions, aluminum melting and steel melting with a little amount on the weld cross-section take place, the wetting and spreading of the liquid aluminum on the steel surface are seen to keep good conditions. The thickness of intermetallic compounds layer is about 9 μm, which are composed of FeAl and Fe3Al. The average shear strength is 27.70 MPa, and the fracture morphology of welding joint is in a ductile and quasi-cleavage mixing mode.

 

Key words: 6016 aluminum alloy; galvanized steel; laser welding; simulation calculation; BP neural network; microstructure; mechanical properties

ISSN 1004-0609
CN 43-1238/TG
CODEN: ZYJXFK

ISSN 1003-6326
CN 43-1239/TG
CODEN: TNMCEW

主管:中国科学技术协会 主办:中国有色金属学会 承办:中南大学
湘ICP备09001153号 版权所有:《中国有色金属学报》编辑部
------------------------------------------------------------------------------------------
地 址:湖南省长沙市岳麓山中南大学内 邮编:410083
电 话:0731-88876765,88877197,88830410   传真:0731-88877197   电子邮箱:f_ysxb@163.com