随着海洋工程领域的扩展与深化，船舶吊运机械臂成为海洋作业的关键技术支持。为了保证机械臂焊接接头的焊接质量，降低海洋作业风险，研究以神经网络为技术支持，设计了点焊接头超声无损检测的缺陷智能识别模型。实验结果表明，研究针对超声无损智能检测模型设计的优化算法，在单峰测试函数上最小值的寻优值为4.804E-11。寻优超体积最大为0.954，与真实前沿解的最小距离为0.203。改进之后的检测模型在F1值上最大可达0.946，损失值最小为0.07，分类检测能力较强。该方法可较好地拟合超声信号特征，有效区分焊接接头的不同缺陷，区分合格焊接与缺陷焊接。研究设计的超声无损缺陷智能检测识别模型有效保证了船舶吊运机械臂的焊接质量，满足海洋工程对船舶吊运机械臂焊接的高要求。
With the expansion and deepening of the field of ocean engineering, ship lifting robotic arms have become a key technical support for ocean operations. In order to ensure the welding quality of the robotic arm's welding joints and reduce the risk of marine operations, a defect intelligent recognition model for spot welding joint ultrasonic non-destructive testing was designed using neural networks as technical support. The experimental results show that the optimization algorithm designed for the ultrasonic non-destructive intelligent testing model has a minimum optimization value of 4.804E-11 on the unimodal testing function. The maximum optimized hyper volume is 0.954, and the minimum distance from the true frontier solution is 0.203. The improved detection model has a maximum F1 value of 0.946 and a minimum loss value of 0.07, indicating strong classification detection ability. This method can well fit the characteristics of ultrasonic signals, effectively distinguish different defects in welded joints, and distinguish between qualified welding and defective welding. The research and design of an ultrasonic non-destructive defect intelligent detection and recognition model effectively ensures the welding quality of ship lifting robotic arms, meeting the high requirements of marine engineering for the welding of ship lifting robotic arms.
2024,46(24): 149-154 收稿日期：2024-5-24
DOI：10.3404/j.issn.1672-7649.2024.24.025
分类号：TP241
基金项目：南通市社会民生科技计划项目（MSZ2023007）
作者简介：季肖枫(1980-),男,硕士,讲师,研究方向为智能化无损检测
参考文献：
[1] 李坤, 任丽. 船舶吊运机械臂海浪环境下的运动补偿控制与模拟[J]. 舰船科学技术, 2023, 45(7): 66-69.
LI K, REN L. Motion compensation control and simulation of ship lifting robotic arm under wave environment[J]. Ship Science and Technology, 2023, 45(7): 66-69.
[2] XIAO W, LI C, FAN L, et al. Dynamic analysis of the lifting arm system in the integrated offshore platform decommissioning equipment in complicated sea states[J]. Processes, 2023, 11(2): 645-668.
[3] 史召一, 曾雯烁, 向瑞豪, 等. 基于内聚力模型的船舶焊接接头裂纹萌生及扩展分析方法[J]. 舰船科学技术, 2023, 45(23): 67-73.
SHI Z Y, ZENG W S, XIANG R H, et al. Crack initiation and extension analysis method of ship welded joints based on cohesion model[J]. Ship Science and Technology, 2023, 45(23): 67-73.
[4] 陈毓, 陈亮, 汪琰, 等. 基于EasyEnsemble和XGBoost算法的焊缝超声波检测结果预测模型[J]. 船舶工程, 2022, 44(4): 134-139.
CHEN Y, CHEN L, WANG Y, et al. Prediction model of weld ultrasonic inspection results based on EasyEnsemble and XGBoost algorithm[J]. Marine Engineering, 2022, 44(4): 134-139.
[5] 郭勇, 王璇, 李菊峰, 等. 电阻点焊焊接质量的超声波无损检测评价方法概述[J]. 焊接技术, 2022, 51(6): 12-18.
GUO Y, WANG X, LI J F, et al. Overview of ultrasonic nondestructive testing methods for evaluating weld quality in resistance spot welding[J]. Welding Technology, 2022, 51(6): 12-18.
[6] DE LEON M, SHIN H S. Prediction of optimum welding parameters for weld-quality characterization in dissimilar ultrasonic-welded Al-to-Cu tabs for Li-ion batteries[J]. Metals and Materials International, 2023, 29(4): 1079-1094.
[7] 张国锋. 基于机器视觉的船体板架结构焊接坏点智能识别方法[J]. 焊接技术, 2023, 52(9): 35-39.
ZHANG G F. Intelligent identification method of welded bad spots in hull plate and frame structure based on machine vision[J]. Welding Technology, 2023, 52(9): 35-39.
[8] LI S, ZOU Y, LAI X, et al. Performance-maximum optimization of the intelligent lifting activities for a polar ship crane through trajectory planning[J]. Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science, 2023, 237(4): 765-781.
[9] TIAN Z. Echo state network based on improved fruit fly optimization algorithm for chaotic time series prediction[J]. Journal of Ambient Intelligence and Humanized Computing, 2022, 13(7): 3483-3502.
