基于多尺度图像特征融合的有机涂层寿命预测研究

doi:10.11902/1005.4537.2025.092

中国腐蚀与防护学报

2025, Vol. 45

Issue (5): 1205-1218 CSTR: 32134.14.1005.4537.2025.092 DOI: 10.11902/1005.4537.2025.092

研究报告

本期目录 | 过刊浏览

基于多尺度图像特征融合的有机涂层寿命预测研究

李婕¹, 孟凡帝¹(

), 孙学思¹, 李佳妮², 陈思涵², 李则蓝², 迟剑宁², 亓海霞³(

), 王福会¹, 刘莉¹

1 东北大学腐蚀与防护中心沈阳 110819
2 东北大学机器人科学与工程学院沈阳 110819
3 中国船舶集团有限公司第七二五研究所海洋腐蚀与防护全国重点实验室厦门 361100

Lifetime Prediction for Organic Coatings via Feature Integration of Multi-Scale Images

LI Jie¹, MENG Fandi¹(

), SUN Xuesi¹, LI Jiani², CHEN Sihan², LI Zelan², CHI Jianning², QI Haixia³(

), WANG Fuhui¹, LIU Li¹

1 Center for Corrosion and Protection, Northeastern University, Shenyang 110819, China
2 Faculty of Robot Science and Engineering, Northeastern University, Shenyang 110819, China
3 National Key Laboratory of Marine Corrosion and Protection, 725th Research Institute of China State Shipbuilding Corporation, Xiamen 361100, China

引用本文:

李婕, 孟凡帝, 孙学思, 李佳妮, 陈思涵, 李则蓝, 迟剑宁, 亓海霞, 王福会, 刘莉. 基于多尺度图像特征融合的有机涂层寿命预测研究[J]. 中国腐蚀与防护学报, 2025, 45(5): 1205-1218.
Jie LI, Fandi MENG, Xuesi SUN, Jiani LI, Sihan CHEN, Zelan LI, Jianning CHI, Haixia QI, Fuhui WANG, Li LIU. Lifetime Prediction for Organic Coatings via Feature Integration of Multi-Scale Images[J]. Journal of Chinese Society for Corrosion and protection, 2025, 45(5): 1205-1218.

全文: PDF(21511 KB) HTML

摘要:

以环氧有机耐磨涂层为研究对象，通过扫描电子显微镜、金相显微镜、激光共聚焦等方法多尺度采集有机涂层的微观形貌，利用基于深度学习的图像识别技术提取图像中的量化参数数据，搭建有机涂层缺陷参数随服役时间的动态演化关系网络以及有机涂层的寿命预测网络模型。结果表明，搭建的演化关系曲线模型以及网络预测模型可较为准确实现对有机涂层的寿命预测研究。

关键词 ：深度学习, 图像识别, 有机涂层, 多尺度特征融合, 寿命预测

Abstract：

Herein, a method for predicting the service life-time of epoxy-based organic anti-abrasive coatings was stablished based on the integrative treatment of the acquired characteristics of microstructure images of multiple scales for organic coatings. Namely, the multiple scale microscopic structural images were collected by means of scanning electron microscopy, metallographic microscopy, laser confocal microscopy and other methods. Then the quantitative parameter data were extracted from the images using image recognition technology based on deep learning. A dynamic evolution relationship model of coating defect parameters with service time and a life prediction network model of organic coatings were constructed. The results indicate that the constructed evolutionary relationship curve model and network prediction model can accurately predict the lifespan of organic coatings.

Key words： deep learning image recognition organic coatings multi-scale feature fusion lifetime prediction

收稿日期: 2025-03-18 32134.14.1005.4537.2025.092

ZTFLH:

TG174

基金资助:国家自然科学基金(52271052);辽宁省自然科学基金(2023-MSBA-043)

通讯作者: 孟凡帝，E-mail：fandimeng@mail.neu.edu.cn，研究方向为海洋环境材料的腐蚀与防护；
亓海霞，E-mail：qihaixia19861222@126.com，研究方向为防腐等功能涂层材料、腐蚀防护设计

Corresponding author: MENG Fandi, E-mail: fandimeng@mail.neu.edu.cn;
QI Haixia, E-mail: qihaixia19861222@126.com

作者简介: 李婕，女，2000年生，硕士生

	服务

	把本文推荐给朋友
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	李婕
	孟凡帝
	孙学思
	李佳妮
	陈思涵
	李则蓝
	迟剑宁
	亓海霞
	王福会
	刘莉
44.8K

链接本文:

https://www.jcscp.org/CN/10.11902/1005.4537.2025.092 或 https://www.jcscp.org/CN/Y2025/V45/I5/1205

图1 涂层样品经不同周期实验后的宏观形貌图像

图2 涂层样品经不同周期实验后的SEM表面形貌与涂层不同类别缺陷图像

图3 构建用于可疑裂纹区域识别的卷积神经网络

图4 磨痕缺陷的数量随实验周期的变化

图5 经不同周期试验后涂层中树脂脱落坑的面积频率分布直方图

图6 涂层经不同周期实验后的金相显微镜基础图像集

图7 不同周次试验后涂层的不同倍数金相照片的平均对比度和平均差异性

图8 不同周期试验后的涂层表面激光共聚焦二维和三维高度图像

图9 CLSM测试的涂层的表面粗糙度参数(Sq和Sa)随浸泡-磨损循环实验周次的变化

图10 MMFCT网络架构图

图11 Transformer网络架构图

图12 交叉尺度特征融合模块

图13 4种模型的损失曲线对比图

表1 3种单尺度深度学习模型与MMFCT融合模型的4种预测性能指标结果对比

图14 4种模型的准确率曲线对比图

图15 4种模型的特异性曲线对比图

图16 4种模型的灵敏度曲线对比图

图17 前20组脱落坑缺陷数量去除离群点后的原始数据

图18 前20组脱落坑缺陷面积占比去除离群点后的原始数据

图19 外延组别的数据检测混淆矩阵

表2 寿命预测模型对外延组别分类的准确度结果

[1]	Zimmermann N, Wang P H. A review of failure modes and fracture analysis of aircraft composite materials [J]. Eng. Fail. Anal., 2020, 115: 104692
[2]	Bakhshan H, Afrouzian A, Ahmadi H, et al. Progressive failure analysis of fiber-reinforced laminated composites containing a hole[J]. Int. J. Damage Mech., 2018, 27: 963
[3]	An D Y, Chen J Y, Liu Z, et al. A method of coating life prediction based on high temperature thermal shock life test and three-dimensional heat transfer analysis [J]. Sci. Rep., 2025, 15: 3029
[4]	Zhang C Z, Fu X Q. Applications and potentials of machine learning in optoelectronic materials research: An overview and perspectives [J]. Chin. Phys. B, 2023, 32: 126103
[5]	Chan C H, Sun M Z, Huang B L. Application of machine learning for advanced material prediction and design [J]. Ecomat, 2022, 4(4): 16
[6]	Xie W B, Zheng H, Li M Y, et al. Membership function-dependent local controller design for t-s fuzzy systems [J]. IEEE. Trans. Syst. Man. Cybern. Syst., 2022, 52(2): 814
[7]	Yoo Y, Baek J G. A novel image feature for the remaining useful lifetime prediction of bearings based on continuous wavelet transform and convolutional neural network [J]. Appl. Sci-Basel., 2018, 8: 1102
[8]	Du J W, Wu F, Yin J C, et al. Polyline simplification based on the artificial neural network with constraints of generalization knowledge [J]. Cartogr. Geogr. Inf. Sci., 2022, 49(4): 313
[9]	Ma H, Liu R, Cui Y, et al. Effect of hydrostatic pressure on corrosion behavior of Cr/GLC laminated coatings [J]. J. Chin. Corros. Prot., 2025, 45: 103
[9]	马宏宇, 刘叡, 崔宇等. 静水压力对Cr/GLC叠层涂层腐蚀行为的影响 [J]. 中国腐蚀与防护学报， 2025, 45: 103 doi: 10.11902/1005.4537.2024.153
[10]	Liu H, Zhang Y F. Image-driven structural steel damage condition assessment method using deep learning algorithm [J]. Measurement, 2019, 133: 168
[11]	Zhang X, Yao J Y, Wu Y L, et al. A method for predicting the creep rupture life of small-sample materials based on parametric models and machine learning models [J]. Materials, 2023, 16: 6804
[12]	Chen Y F, Meng F D, Liu L, et al. Life prediction of organic coatings based on grey system theory in coupled deep-sea pressure-flow velocity environment [J]. Equip. Environ. Eng., 2023, 20(2):64
[12]	陈宇凡, 孟凡帝, 刘莉等. 深海压力-流速耦合环境下基于灰色系统理论的有机涂层寿命预测研究 [J]. 装备环境工程, 2023, 20(2): 64
[13]	Geng G Q, Lin J, Liu L J, et al. Prediction method for life of anti-corrosion coating for steel bridge [J]. Natur. Sci. Ed., 2006, (5): 43
[13]	耿刚强, 林杰, 刘来君等. 钢桥防腐蚀涂层寿命的预测方法 [J]. 长安大学学报(自然科学版), 2006, (5): 43
[14]	Choudhary K, Decost B, Chen C, et al. Recent advances and applications of deep learning methods in materials science [J]. NPJ Comput. Mater., 2022, 8(1): 548
[15]	Zarkevich N A. Theoretical and computational methods for accelerated materials discovery [J]. Mod. Phys. Lett. B, 2021, 35(12): 213
[16]	Zhai H, Zhai Y J. Optimization design of ferry material performance test system based on artificial intelligence [J]. J. Nanomater., 2022, 2022: 1155
[17]	Yu W J, Wang T C, Zhao D Y, et al. Research on life prediction model of closed corrosion-resistant coatings [J]. J.Chin.Soc.Corros.Prot., 2024, 44: 1617
[17]	禹文娟, 王天丛, 赵东杨等. 封闭型耐蚀涂层的寿命预测模型研究 [J]. 中国腐蚀与防护学报， 2024, 44: 1617 doi: 10.11902/1005.4537.2023.383
[18]	Lopes C F, Machado G S D, Mendes P S N, et al. Analysis of copper and zinc alloy surface by exposure to alcohol aqueous solutions and sugarcane liquor [J]. J. Mater. Res. Technol., 2020, 9(2): 2545
[19]	Da Silva P C, Junior C F L, Huguenin J a O, et al. Investigation of copper and zinc alloy surface exposed to corrosion environment by digital image processing [J]. J. Mater. Res. Technol., 2023, 24: 9743
[20]	Yao Y, Yang Y, Wang Y P, et al. Artificial intelligence-based hull structural plate corrosion damage detection and recognition using convolutional neural network [J]. Appl. Oce. Res., 2019, 90: 101823
[21]	Pan J, Liu F, Feng J, et al. Accurate recognition of micromorphology images of epoxy coatings for deep-sea environments based on a deep learning super-resolution method [J]. Corros. Commun., 2025, 19: 14

[1]	郭玉杰, 李艳辉, 夏大海, 胡文彬. 腐蚀电化学阻抗谱的数据解析与物理模型研究进展[J]. 中国腐蚀与防护学报, 2025, 45(5): 1143-1160.
[2]	申志远, 单广斌, 陈闽东, 刘媛双. 用于腐蚀监测的电化学噪声信号识别方法[J]. 中国腐蚀与防护学报, 2025, 45(5): 1433-1440.
[3]	陈丽娟, 晁刘伟, 赵景茂. CeO₂@Zr-MOF复合材料的制备及其对环氧涂层保护性能的提升作用[J]. 中国腐蚀与防护学报, 2025, 45(3): 664-674.
[4]	刘明, 张连栋, 孙志华, 高蒙, 闫巍, 赵明亮. 典型氟橡胶密封材料热空气老化试验研究[J]. 中国腐蚀与防护学报, 2025, 45(1): 137-147.
[5]	禹文娟, 王天丛, 赵东杨, 向雪云, 吴航, 王文. 封闭型耐蚀涂层的寿命预测模型研究[J]. 中国腐蚀与防护学报, 2024, 44(6): 1617-1624.
[6]	李卓玄, 曹艳辉, 李崇杰, 李辉, 张小明, 雍兴跃. 耦接件涂层失效程度与其力学损伤之间的关系[J]. 中国腐蚀与防护学报, 2024, 44(3): 679-690.
[7]	王凯, 李晨沛, 卢金玲, 王振江, 王维. NiCoCrFeNb_0.45 共晶高熵合金在水力机械中的抗空蚀性能研究[J]. 中国腐蚀与防护学报, 2023, 43(5): 1079-1086.
[8]	孟凡帝, 高浩东, 刘莉, 崔宇, 刘叡, 王福会. 适用于深海压力-流体耦合环境的玄武岩有机防腐涂层的制备及性能研究[J]. 中国腐蚀与防护学报, 2023, 43(4): 704-712.
[9]	王通, 王巍. 聚二甲基硅氧烷涂层自修复过程中的弛豫时间分布研究[J]. 中国腐蚀与防护学报, 2023, 43(2): 337-344.
[10]	陈异凡, 孟凡帝, 曲优异, 方芷晴, 刘莉, 王福会. 超疏水聚苯胺胶囊的一步可控合成及其对有机涂层防腐性能的影响[J]. 中国腐蚀与防护学报, 2023, 43(2): 345-351.
[11]	王明好, 王欢, 刘叡, 孟凡帝, 刘莉, 王福会. 基于深度学习方法的N5/NiCrAlY涂层图像识别的研究[J]. 中国腐蚀与防护学报, 2022, 42(4): 583-589.
[12]	栾浩, 孟凡帝, 刘莉, 崔宇, 刘叡, 郑宏鹏, 王福会. 间苯二胺-氧化石墨烯/有机涂层的制备及防腐性能研究[J]. 中国腐蚀与防护学报, 2021, 41(2): 161-168.
[13]	曹京宜, 王智峤, 李亮, 孟凡帝, 刘莉, 王福会. 深海压力交变加速条件下改性石墨烯有机涂层的失效机制[J]. 中国腐蚀与防护学报, 2020, 40(2): 139-145.
[14]	邓三喜, 闫小宇, 柴柯, 吴进怡, 史洪微. 假单胞菌对聚硅氧烷树脂清漆涂层分解及防腐蚀行为的影响[J]. 中国腐蚀与防护学报, 2018, 38(4): 326-332.
[15]	桂琪, 郑大江, 宋光铃. 醇酸清漆保护性的电化学加速评价[J]. 中国腐蚀与防护学报, 2018, 38(3): 274-282.

Viewed

Full text

Abstract

Cited

Shared

Discussed