机器学习在磁记忆无损检测领域的应用及展望

doi:10.3969/j.issn.1674-0696.2022.08.09

摘要/Abstract

摘要： 磁记忆检测作为一种新兴的无损检测技术，具有快速、便捷地检测应力集中和微观缺陷的优点，而机器学习具有强大的学习能力和自适应能力，适用于磁记忆检测产生的大量非线性数据的处理。主要综述了机器学习算法在磁记忆无损检测领域的应用现状，考虑将机器学习应用于桥梁内部钢筋损伤的磁记忆检测当中，最终展望了机器学习在此领域的发展趋势。结果表明：以支持向量机、神经网络、聚类算法为代表的各种机器学习算法目前在磁记忆检测中得到了广泛应用，主要用于各类钢试件缺陷的损伤等级评估和缺陷尺寸反演，在桥梁钢筋损伤的磁记忆检测中尚无相关应用；单一算法具有较大的局限性，多种算法结合可以提高预测准确率；要实现磁记忆在桥梁结构内部钢筋损伤检测中的突破发展，需进一步考虑复杂结构和检测环境，结合机器学习算法建立各类影响因素与磁信号之间的关系模型，机器学习中的回归算法也可进一步应用到缺陷程度的评估当中。

关键词: 桥梁工程；机器学习；磁记忆；无损检测；应用；展望

Abstract: As a newly emerged nondestructive testing technology, magnetic memory testing has the advantages of rapid and convenient detection of stress concentration and micro defects. Machine learning has strong learning ability and self-adaptive ability, which is suitable for processing large amount of non-linear data generated by magnetic memory detection. This article mainly reviewed the application status of machine learning algorithms in the field of magnetic memory nondestructive testing and considered the application of machine learning in magnetic memory detection of reinforcement damage inside bridges. Finally, the development trend of machine learning in this field was prospected. The results show that various machine learning algorithms represented by SVM, ANN and clustering have been widely used in magnetic memory testing, which are mainly used for damage grade evaluation and defect size inversion of steel specimens and have no relevant application in magnetic memory detection of bridge reinforcement damage. A single algorithm has great limitations, and the combination of multiple algorithms can improve prediction accuracy. To achieve the breakthrough development of magnetic memory in the detection of reinforcement damage inside bridges, it is necessary to consider the complex structure and detection environment. In the future, machine learning algorithm will be used to establish the relationship model between various influencing factors and magnetic signals. The regression algorithm in machine learning can also be further applied to the evaluation of defect degree.

Key words: bridge engineering; machine learning; magnetic memory; non-destructive testing; application; prospect

中图分类号:

U449.7
TP399

杨茂1,2，张洪1，周建庭1，刘人铭1，陈军1. 机器学习在磁记忆无损检测领域的应用及展望[J]. 重庆交通大学学报（自然科学版）, 2022, 41(08): 58-66.

YANG Mao1,2，ZHANG Hong1, ZHOU Jianting1, LIU Renming1, CHEN Jun1. Application and Prospect of Machine Learning in the Field of Magnetic Memory Nondestructive Testing[J]. Journal of Chongqing Jiaotong University(Natural Science), 2022, 41(08): 58-66.

参考文献

［1］ DUBOV A A. A study of metal properties using the method of magnetic memory［J］. Metal Science and Heat Treatment, 1997, 39(9): 401-405.
［2］董丽虹, 徐滨士, 朱子新, 等. 铁磁材料磁记忆检测技术的研究现状［J］. 新技术新工艺, 2005(9):24-27.
DONG Lihong, XU Binshi, ZHU Zixin, et al. The present state of metal magnetic memory testing for ferromagnetic materials ［J］. New Technology and New Process,
2005(9): 24-27.
［3］王威, 易术春, 苏三庆, 等. 金属磁记忆无损检测的研究现状和关键问题［J］. 中国公路学报, 2019, 32(9): 1-21.
WANG Wei, YI Shuchun, SU Sanqing, et al. Research status and critical problems of metal magnetic memory testing［J］. China Journal of Highway and Transport,
2019,32(9):1-21.
［4］ WANG Zhengdao, GU Yuanxun, WA Yuesheng. A review of three magnetic NDT technologies［J］. Journal of Magnetism and Magnetic Materials, 2012, 324(4):
382-388.
［5］杨茂, 周建庭, 张洪, 等. 混凝土内部钢筋锈蚀的磁记忆检测［J］. 建筑材料学报, 2018, 21(2): 345-350.
YANG Mao, ZHOU Jianting, ZHANG Hong, et al. Magnetic memory detection of rebar corrosion in concrete［J］. Journal of Building Materials, 2018,21(2):345-350.
［6］ DUBOV A A. Diagnostics of steam turbine disks using the metal magnetic memory method［J］. Thermal Engineering, 2010, 57(1): 16-21.
［7］ KOLOKOLNIKOV S M, DUBOV A A, MARCHENKOV A Y. Determination of mechanical properties of metal of welded joints by strength parameters in the stress
concentration zones detected by the metal magnetic memory method［J］. Welding in the World, 2014, 58(5): 699-706.
［8］ DUBOV A, KOLOKOLNIKOV S. The metal magnetic memory me-thod application for online monitoring of damage development in steel pipes and welded joints
specimens［J］. Welding in the World, 2013, 57(1): 123-136.
［9］李壮年, 储满生, 柳政根, 等. 基于机器学习和遗传算法的高炉参数预测与优化［J］. 东北大学学报(自然科学版), 2020,41(9):1262-1267.
LI Zhuangnian, CHU Mansheng, LIU Zhenggen, et al. Prediction and optimization of blast furnace parameters based on machine learning and genetic algorithm［J］.
Journal of Northeastern University (Natural Science Edition), 2020, 41(9): 1262-1267.
［10］刘留, 张建华, 樊圆圆, 等. 机器学习在信道建模中的应用综述［J］. 通信学报, 2020, 41(8): 1-20.
LIU Liu, ZHANG Jianhua, FAN Yuanyuan, et al. Survey of application of machine learning in wireless channel modeling ［J］. Journal on Communications, 2020, 41(8): 1-
20.
［11］陈海昕, 邓凯文, 李润泽. 机器学习技术在气动优化中的应用［J］. 航空学报, 2019,40(1):47-63.
CHEN Haixin, DENG Kaiwen, LI Runze. Application of machine learning technology in aerodynamic optimization［J］. Acta Aeronautica Sinica, 2019, 40(1): 47-63.
［12］ ZHAO Rui, YAN Ruiqiang, CHEN Zhenghua, et al. Deep learning and its applications to machine health monitoring［J］. Mechanical Systems and Signal Processing,
2019,115: 213-237.
［13］ HOANG D, KANG H. A survey on deep learning based bearing fault diagnosis［J］. Neurocomputing, 2019, 335: 327-335.
［14］ KHAN S, YAIRI T. A review on the application of deep learning in system health management［J］. Mechanical Systems and Signal Processing, 2018, 107: 241-265.
［15］米晓希, 汤爱涛, 朱雨晨, 等. 机器学习技术在材料科学研究中的应用进展［J］. 材料导报, 2021,35(15):15115-15124.
MI Xiaoxi, TANG Aitao, ZHU Yuchen, et al. Machine learning: A potential powerful tool for materials science research ［J］. Materials Review, 2021,35(15): 15115-15124.
［16］邢海燕, 葛桦, 秦萍, 等. 基于遗传神经网络的焊缝缺陷等级磁记忆定量化研究［J］. 材料科学与工艺, 2015,23(2):33-38.
XING Haiyan, GE Hua, QIN Ping, et al. Quantitative research on magnetic memory of weld defect grade based on genetic neural network［J］. Materials Science and
Technology, 2015,23(2): 33-38.
［17］秦萍. 基于优化算法的焊缝缺陷等级磁记忆定量化研究［D］. 大庆: 东北石油大学, 2014:33-35.
QIN Ping. MMM Quantifying of Welded Joint Defect Levels Based on BP Neural Network Optimized by Genetic Algorithm［D］. Daqing: Northeast Petroleum University,
2014:33-35.
［18］邢海燕, 孙晓军, 王犇, 等. 基于模糊加权马尔科夫链的焊缝隐性损伤磁记忆特征参数定量预测［J］. 机械工程学报, 2017,53(12):70-77.
XING Haiyan, SUN Xiaojun, WANG Ben, et al. Quantitative MMM characteristic parameter prediction for weld hidden damage status based on the fuzzy weighted markov
chain ［J］. Journal of Mechanical Engineering, 2017,53(12): 70-77.
［19］邢海燕, 喻正帅, 李雪峰, 等. 基于模糊c均值聚类算法的焊缝缺陷等级磁记忆定量识别［J］. 压力容器, 2018,35(6):57-63.
XING Haiyan, YU Zhengshuai, LI Xuefeng, et al. Quantitative metal magnetic memory identification of weld defect levels based on fuzzy c-means clustering algorithm ［J
］. Pressure Vessel, 2018,35(6):57-63.
［20］喻正帅. 基于模糊聚类分析的磁记忆信号临界特征提取研究［D］. 大庆：东北石油大学, 2018:30-32
YU Zhengshuai. Research on Extracting Critical Features of Magnetic Memory Signals Based on Fuzzy Clustering Analysis［D］.Daqing: Northeast Petroleum University,
2018: 30-32.
［21］邢海燕, 陈玉环, 李雪峰, 等. 基于动态免疫模糊聚类的金属焊缝缺陷等级磁记忆识别模型［J］. 仪器仪表学报, 2019,40(11):225-232.
XING Haiyan, CHEN Yuhuan, LI Xuefeng, et al. Magnetic memory identification model of mental weld defect levels based on dynamic immune fuzzy clustering ［J］.
Chinese Journal of Scientific Instrument, 2019, 40(11): 225-232.
［22］易方, 李著信, 吕宏庆, 等. 基于模糊核支持向量机的管道磁记忆检测缺陷识别［J］. 石油学报, 2010,31(5):863-866, 870.
YI Fang, LI Zhuxin, LYU Hongqing, et al. Defect recognition by metal magnetic memory detection of pipelines based on the fuzzy kernel function SVM ［J］. Acta Petrolei
Sinica, 2010, 31(5): 863-866,870.
［23］ LIU Zhilin, LIU Lutao, ZHANG Jun. Signal feature extraction and quantitative evaluation of metal magnetic memory testing for oil well casing based on data
preprocessing technique［J］. Abstract and Applied Analysis, 2014:1-9.
［24］ GONG Lihong, LI Zhuxin, ZHANG Zhen. Diagnosis model of pipeline cracks according to metal magnetic memory signals based on adaptive genetic algorithm and
support vector machine［J］. The Open Mechanical Engineering Journal, 2015, 9(1): 1076-1080.
［25］龚利红, 李著信, 许红, 等. 基于感知器神经网络的金属磁记忆检测管道缺陷分析［J］. 机床与液压, 2013,41(9):186-188.
GONG Lihong, LI Zhuxin, XU Hong, et al. Analysis of pipeline defects by metal magnetic memory detection based on perceptron neural network［J］. Machine Tool &
Hydraulics, 2013, 41(9): 186-188.
［26］刘书俊, 蒋明, 张伟明, 等. 基于BP神经网络的油气管道缺陷磁记忆检测［J］. 无损检测, 2015,37(7):25-28.
LIU Shujun, JIANG Ming, ZHANG Weiming, et al. Magnetic memory testing on oil and gas pipeline based on BP neural network［J］. Nondestructive Testing, 2015, 37(7):
25-28.
［27］李远利, 李著信, 刘书俊. 模糊免疫算法及其在金属磁记忆检测中的应用［J］. 微型机与应用, 2012,31(6):69-71.
LI Yuanli, LI Zhuxin, LIU Shujun. Application of fuzzy-immune algorithm in metal magnetic memory testing signal analysis［J］. Microcomputers and Applications, 2012,
31(6): 69-71.
［28］李效露. 基于小波奇异性和神经网络的钢绳芯输送带故障诊断方法的研究［D］. 太原：太原理工大学, 2013：65-71.
LI Xiaolu. Research of Fault Diagnosis Method of Steel Cord Conveyor Belt Based on Wavelet Singularity and Neural Network［D］. Taiyuan: Taiyuan University of
Technology, 2013: 65-71.
［29］王慧鹏, 董丽虹, 董世运, 等. 基于磁记忆的应力集中神经网络识别［J］. 理化检验-物理分册, 2013,49(9):576-579.
WANG Huipeng, DONG Lihong, DONG Shiyun, et al. Neural network recognition of stress concentration based on magnetic memory testing［J］. Physical Testing and
Chemical Analysis Part A:Physical Testing, 2013, 49(9): 576-579.
［30］ SHU Di, YIN Ling, BU Jin, et al. Application of a combined metal magnetic memory-magnetic Barkhausen noise technique for on-site detection of the stress-free
temperature of a continuous welded rail［J］. Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit, 2014, 230(3): 774-783.
［31］邢海燕, 党永斌, 王犇, 等. 基于K-近邻隶属度模糊支持向量机的再造抽油杆损伤等级磁记忆定量识别［J］. 石油学报, 2015,36(11):1427-1432, 1456.
XING Haiyan, DANG Yongbin, WANG Ben, et al. Quantitative MMM identification of damage levels based on KNN FSVM for remanufactured sucker rod［J］. Acta
Petrolei Sinica, 2015, 36(11): 1427-1432, 1456.
［32］李思岐, 俞洋, 党永斌, 等. 基于改进的支持向量回归机算法的磁记忆定量化缺陷反演［J］. 工程科学学报, 2018,40(9):1123-1130.
LI Siqi, YU Yang, DANG Yongbin, et al. Metal magnetic memory quantitative inversion of defects based on optimized support vector machine regression［J］. Journal of
Engineering Science, 2018, 40(9): 1123-1130.
［33］李立刚, 万勇, 王宇, 等. 基于支持向量机和磁记忆技术的管道缺陷深度的定量化反演研究［J］. 腐蚀与防护, 2020,41(1):29-34,40.
LI Ligang, WAN Yong, WANG Yu, et al. Quantitative inversion of pipeline defect depth based on support vector machine and magnetic memory technology［J］.
Corrosion and Protection, 2020, 41(1): 29-34,40.
［34］王帅, 黄海鸿, 韩刚, 等. 基于PCA与GA-BP神经网络的磁记忆信号定量评价［J］. 电子测量与仪器学报, 2018,32(10):190-196.
WANG Shuai, HUANG Haihong, HAN Gang, et al. Quantitative evaluation of magnetic memory signal based on PCA & GA-BP neural network［J］. Journal of Electronic
Measurement and Instrument, 2018, 32(10): 190-196.
［35］王宇, 万勇, 杨勇, 等. 基于极限学习机及磁记忆技术的管道缺陷分类方法研究［J］. 油气田地面工程, 2019,38(7):98-103.
WANG Yu, WAN Yong, YANG Yong, et al. Classification method research on pipeline defects based on extreme learning machine and magnetic memory technology［J］.
Oil and Gas Field Surface Engineering, 2019, 38(7): 98-103.
［36］史小东, 樊建春, 周威, 等. 基于BP神经网络管道磁记忆检测模式识别［J］. 石油机械, 2020,48(6):111-117.
SHI Xiaodong, FAN Jianchun, ZHOU Wei, et al. Pattern recogni-tion of pipeline magnetic memory inspection based on BP neural network ［J］. Petroleum Machinery,
2020, 48(6): 111-117.
［37］ GAO Yatian, LENG Jiancheng, LI Siqi. Residual life prediction me-thod for remanufacturing sucker rods based on magnetic memory testing and a support vector
machine model［J］. Insight (Northampton), 2019, 61(1): 44-50.
［38］周宏强, 黄玲玲, 王涌天. 深度学习算法及其在光学的应用［J］. 红外与激光工程, 2019, 48(12): 299-318.
ZHOU Hongqiang, HUANG Lingling, WANG Yongtian. Deep learning algorithm and its application in optics［J］. Infrared and Laser Engineering, 2019, 48(12): 299-318.
［39］牛程程, 李少波, 胡建军, 等. 机器学习在材料信息学中的应用综述［J］. 材料导报, 2020,34(23):23100-23108.
NIU Chengcheng, LI Shaobo, HU Jianjun, et al. Application of machine learning in material informatics: A survey［J］. Materials Review, 2020, 34(23): 23100-23108.

[1]	梁宗保1，柴洁1，纳守勇2，马天立1，唐玉1. 基于深度学习的桥梁健康监测数据有效性分析[J]. 重庆交通大学学报（自然科学版）, 2021, 40(03): 78-83.
[2]	李修云，黄硕，杨文伟，谭超英. 面向桥梁结构状态预测的ARMA-GM组合时序模型研究[J]. 重庆交通大学学报（自然科学版）, 2016, 35(4): 6-9.