[1]张文静,杨学坤,李明亮,等.doi: 10.3969/j.issn.1001-3849.2026.03.009热浸镀Zn-Al-Mg合金镀层表面热疲劳裂纹自动识别[J].电镀与精饰,2026,(03):70-76.
　ZHANG Wenjing,YANG Xuekun,LI Mingliang,et al.Automatic identification of thermal fatigue cracks on the surface of hot dip Zn-Al-Mg alloy coating[J].Plating & Finishing,2026,(03):70-76.
点击复制

doi: 10.3969/j.issn.1001-3849.2026.03.009热浸镀Zn-Al-Mg合金镀层表面热疲劳裂纹自动识别()

《电镀与精饰》[ISSN:1001-3849/CN:12-1096/TG]

卷:
期数:
2026年03
页码:
70-76
栏目:
出版日期:
2026-03-31

文章信息/Info

Title:
Automatic identification of thermal fatigue cracks on the surface of hot dip Zn-Al-Mg alloy coating
作者:
张文静1; 杨学坤1; 李明亮2; 刘 茵1
(1. 北京农业职业学院 智慧农业工程学院,北京 102442 ;2. 河北地质大学 信息中心,河北 石家庄 050031)
Author(s):
ZHANG Wenjing1;  YANG Xuekun1;  LI Mingliang2;  LIU Yin1
(1. College of Smart Agriculture Engineering, Beijing Vocational College of Agriculture, Beijing 102442, China; 2. Information Center, Hebei GEO University, Shijiazhuang 050031, China)
关键词:
热浸镀; Zn-Al-Mg合金镀层; 热疲劳裂纹; 自动识别; 卷积神经网络
Keywords:
hot dip plating;  Zn-Al-Mg alloy coating;  thermal fatigue crack;  automatic identification;  convolutional neural network
分类号:
TP391 TQ153.2
文献标志码:
A
摘要:
不同合金结构元素在受到周期性变化的影响下其热应力变化也不同，热疲劳会导致合金开裂产生裂纹，而小于结构元素的干扰会增加表面镀层图像的噪点，灰度特征容易与背景融合，降低了对裂纹定位的精度，热疲劳裂纹识别效果不佳。本文针对热浸镀Zn-Al-Mg合金镀层表面热疲劳裂纹自动识别方法展开研究。考虑到热浸镀Zn-Al-Mg合金结构元素的热应力变化，对镀层图像进行腐蚀和膨胀运算，并基于数学形态滤波，完成镀层表面图像滤波去噪处理；构建卷积神经网络模型，添加空间注意力机制以及信道注意力机制，提取热疲劳裂纹区域特征图，与背景分离；在训练模型过程中引入损失函数，提高对裂纹定位的精度，并将裂纹识别转换成二分类问题，实现合金镀层表面热疲劳裂纹自动识别。通过实验证明：利用所提方法进行热疲劳裂纹自动识别，峰值信噪比保持在25 dB以上，不仅能够准确捕捉裂纹的整体形态，而且就连极其微小的细节部分也能被有效识别，平均计算量为2.2 GFLOPs，平均参数量为1 435 106.6 M，能够以最快的速度完成裂纹识别任务，应用效果较好。
Abstract:
Under the influence of periodic changes, the thermal stress changes of different alloy structural elements are also different. Thermal fatigue can lead to alloy cracking and crack formation, while interference smaller than structural elements can increase the noise of surface coating images. Gray scale features are easily fused with the background, reducing the accuracy of crack localization, and the recognition effect of thermal fatigue cracks is poor. Therefore, this article focuses on the research of automatic identification method for thermal fatigue cracks on the surface of hot-dip Zn-Al-Mg alloy coating. Considering the thermal stress changes of structural elements in hot-dip Zn-Al-Mg alloy coating, corrosion and expansion operations were performed on the coating image. Based on mathematical morphology filtering, the surface image of the coating was filtered and denoised. The feature map of the thermal fatigue crack area was extracted and separated from the background using a convolutional neural network model equipped with spatial and channel attention mechanisms. Finally, a loss function is introduced during the training process to improve the accuracy of crack localization, and crack identification is transformed into a binary classification problem, achieving automatic identification of thermal fatigue cracks on the surface of alloy coatings. Experimental results show that using the proposed method for automatic identification of thermal fatigue cracks can maintain a peak signal-to-noise ratio of over 25 dB, accurately capturing the overall shape of the crack and effectively identifying even extremely small details. The average computational cost is 2.2 GFLOPs and the average parameter size is 1 435 106.6 M, enabling the fastest completion of crack identification tasks and achieving better application

参考文献/References:

[1].姚良少. 强力磁铁在碱性锌锰电池生产线上的应用[J]. 电池, 2024(2): 227-229.
[2].李涛, 柳伟, 李清萍. Zn2+掺杂硫化物/氢氧化物异质结构的性能[J]. 电池, 2024(3): 297-302.
[3].江守燕, 邓王涛, 孙立国, 等. 数据驱动的半无限介质裂纹识别模型研究[J]. 力学学报, 2024, 56(6): 1727-1739.
[4].王慧鹏, 翟思聪, 杨洁, 等. 基于激光热成像的铝合金表面裂纹宽度定量检测[J]. 红外与激光工程, 2023, 52(9): 154-163.
[5].肖奇, 冯健, 汪刚, 等. 基于多频平衡电磁技术的管道内外缺陷检测与全角度表面裂纹识别方法[J]. 中国电机工程学报, 2023, 43(14): 5703-5713.
[6].劳武略, 崔闯, 张登科, 等. 基于计算机视觉的钢桥面板裂纹识别方法[J]. 中国公路学报, 2023, 36(3): 188-201.
[7].HOLZMON O, ROACHE D C, PRICE M C, et al. Enhancing crack detection in critical structures using machine learning and 3D digital Image correlation[J]. Experimental Mechanics, 2024, 64(8): 1369-1380.
[8].郑杰, 权伟, 刘洋, 等. Cr/Zr涂层环向压缩表面裂纹识别研究[J]. 机械工程学报, 2024, 60(8): 1-10.
[9].陈博, 汪学清, 董泽, 等. 基于DIC与声发射技术的裂纹识别与损伤变量分析[J]. 矿业研究与开发, 2024, 44 (10): 150-158.
[10].王亚飞, 杨浩哲, 勾红叶, 等. 基于深度学习的正交异性钢桥面板疲劳裂纹智能识别方法研究[J]. 桥梁建设, 2024, 54(1): 68-74.
[11].胡子健, 梁森, 孙瑞骏, 等. 嵌入式光纤阵列复合材料裂纹损伤在线识别研究[J]. 化工新型材料, 2024, 52(9): 176-179, 188.
[12].郝柏桥, 范玉刚, 宋执环. 基于深度迁移学习的脉冲涡流热成像裂纹缺陷检测[J]. 光学学报, 2023, 43(4): 146-154.
[13].吕亮亮, 张伟斌, 李公平, 等. 基于深度学习的PBX裂纹像素级识别方法[J]. 含能材料, 2024, 32(5): 545-553.
[14].欧阳煜, 夏登科, 楚鹏辉. 基于裂纹诱导弦挠度的梁裂纹损伤识别[J]. 上海大学学报(自然科学版), 2023, 29(6): 1042-1052.
[15].席特, 蒋勇, 杨飞, 等. SA508-3/690异种金属焊接接头显微组织及裂纹形成机理[J]. 热加工工艺, 2024, 53(13): 35-39.
[16].卢楠楠, 郭以沫, 杨树林, 等. 激光增材修复单晶高温合金的热裂纹形成机制[J]. 金属学报, 2023, 59(9): 1243-1252.
[17].海良豪, 赵纪元, 王琛玮, 等. 经验小波变换的激光超声定量识别表面裂纹[J]. 压电与声光, 2024, 46(5): 776-786.
[18].李坤鹏, 李彪, 张彦军, 等. 基于应变监控数据的金属结构疲劳裂纹量化模型研究[J]. 西北工业大学学报, 2023, 41(5): 932-941.
[19].肖孝鹏, 董文超, 李殿中, 等. 温度对Super304H焊缝金属再热裂纹敏感性的影响[J]. 稀有金属材料与工程, 2023, 52(7): 2444-2452.

更新日期/Last Update: 2026-03-11