不锈钢激光焊接头超声导波检测与质量评价技术  

Research on Ultrasonic Guided Wave Detection and Quality Evaluation Technology for Stainless Steel Laser Welding Head

作　　者：赵雪山 段珍珍[2] 周广浩 郑月生 谷晓鹏[3] ZHAO Xueshan;DUAN Zhenzhen;ZHOU Guanghao;ZHENG Yuesheng;GU Xiaopeng(CRRC Chang Chun Railway Vehicles Co.,Ltd.,Changchun 130062,China;Changchun Institute Of Technology,Changchun 130103;Jilin University,Changchun 130022)

机构地区：[1]中车长春轨道客车股份有限公司,吉林长春130062 [2]长春工程学院,吉林长春130103 [3]吉林大学,吉林长春130022

出　　处：《电焊机》2024年第5期142-147,159,共7页Electric Welding Machine

基　　金：吉林省科技发展计划项目(20220201039GX)。

摘　　要：半熔透型搭接激光焊工艺质量控制难度大,如何有效保证板层间的熔合宽度是开发和应用半熔透型搭接激光焊技术的关键。为此,开展薄板焊接接头超声导波检测方法研究,研发接头内部熔合宽度检测与评估系统,通过对超声导波信号时、频域特征分析,确定能够表现接头内部熔合状态的信号特征值,并以此建立能够有效区分不同焊接质量的评价模型,从而实现不锈钢半熔透型搭接激光焊接头内部熔宽的在线检测。研究结果表明,以超声导波时域信号上最大归一化幅值、频域上第一波峰和第二波峰的峰值为输入信息,采用BP神经网络算法进行分类预测,能够有效识别未熔合、熔合宽度不足、熔合宽度合格等连接状态,预测准确率可达94.4%。The laser welding process with a semi-penetration lap joint,due to its advantages of high efficiency and aesthet‐ics,is currently highly valued as a manufacturing technology for the new generation of stainless steel subway body.How‐ever,the quality control of this welding process is difficult,and how to effectively ensure the fusion width between the lay‐ers is the key to developing and applying laser welding technology with a semi-penetration lap joint.To this end,research was conducted on the ultrasonic guided wave detection method for thin plate welded joints,and a system for detecting and evaluating the internal fusion width of joints was developed.By analyzing the time-frequency characteristics of ultrasonic guided wave signals,the signal characteristic values that can reflect the internal fusion state of joints were determined.Based on this,an evaluation model that can effectively distinguish different welding qualities was established,thus achiev‐ing online detection of the internal fusion width of semi-penetration laser welded lap joints for stainless steel.The research results show that using the maximum normalized amplitude of the ultrasonic guided wave time-domain signal and the peak values of the first and second peaks in the frequency domain as input information,and using the BP neural network algo‐rithm for classification prediction can effectively identify different connection states such as incomplete fusion,insufficient fusion width,and qualified fusion width,with a prediction accuracy of 94.4%.

关 键 词：激光焊接头 超声导波 无损检测 质量评价 

分 类 号：TG441.7[金属学及工艺—焊接]