首页 > 论文 > 红外技术 > 41卷 > 12期(pp:1141-1145)

涡流热成像隐马尔科夫评估方法及应用

The Evaluation Method and Application of Hidden Markov in Eddy Current Thermal Imaging

  • 摘要
  • 论文信息
  • 参考文献
  • 被引情况
  • PDF全文
分享:

摘要

电涡流脉冲热成像(Eddy current pulsed thermography,ECPT)技术是一种集成了电、磁、热等多物理效应的新兴无损检测技术,具有单次检测范围大、无交互、检测速度快等特点。传统基于特定时刻的单帧热图像分析,很难直接区分不同区域的热分布,无法建立有效的数学、物理模型,对微小缺陷与早期疲劳损伤难以进行完整、有效的检测与评估。本文利用多帧热图像序列构造出了不同阶段的多维观测矩阵,通过拉普拉斯特征映射( Laplacian eigenmap,LE)方法对高维观测数据集进行降维处理。利用 COMSOL仿真分析了隐马尔科夫模型( hidden Markov model,HMM)追踪、评估材料特性的有效性,最后通过 HMM对降维后的齿轮疲劳数据进行学习、评估。通过不同阶段的齿轮疲劳实验验证,基于隐马尔科夫模型与电涡流脉冲热成像的技术可以很好地评估齿轮在不同阶段的疲劳损伤状态。

Abstract

Eddy current pulsed thermography (ECPT) is a new nondestructive testing technology that integrates the many physical effects of electricity, magnetism, and heat. ECPT has characteristics such as a large single detection range, non-interaction, and high detection efficiency. Traditional single-frame thermal image analysis based on a specific moment makes it difficult to directly separate thermal distributions of different regions, thereby making it impossible to establish effective mathematical and physical models. Besides, it is challenging to conduct complete and effective detection and evaluation of micro defects and early fatigue damage. In this work, multi-frame thermal image sequences were used to construct multi-dimensional observation matrices at different stages. A Laplacian eigenmap (LE) was adopted to reduce the dimensionality of the observation data set. COMSOL software was used to simulate and analyze the effectiveness of hidden Markov model(HMM) tracking and to evaluate material characteristics. Finally, HMM was used to train and evaluate gear fatigue data after dimension reduction. It can be verified by gear fatigue experiments that the combination of the hidden Markov model and eddy current pulse thermal imaging technology can be used to effectively evaluate the fatigue damage of a gear at different stages.

Newport宣传-MKS新实验室计划
补充资料

所属栏目:无损检测

基金项目:重庆市科技重大主题专项( cstc2018jszx-cyztzxX0032)

收稿日期:2019-07-04

修改稿日期:2019-11-28

网络出版日期:--

作者单位    点击查看

尹爱军:重庆大学机械工程学院,重庆 400044重庆大学机械传动国家重点实验室,重庆 400044
姚文杰:重庆大学机械工程学院,重庆 400044重庆大学机械传动国家重点实验室,重庆 400044

联系人作者:尹爱军(aijun.yin@cqu.edu.cn)

备注:尹爱军( 1978-),男,重庆大学教授、博士生导师。专业方向包括无损检测、故障诊断、大数据分析与智能测试、机电一体化等。 E-mail: aijun.yin@cqu.edu.cn。

【1】袁丽华 , 汪江飞 , 朱争光 , 等. 材料表面裂纹的红外热像显微检测 [J].红外技术, 2018, 40(6): 612-617. YUAN Lihua, WANG Jiangfei, ZHU Zhengguang, et al. Detection of material surface cracks by infrared thermal imaging microscope[J]. Infrared Technology, 2018, 40(6): 612-617.

【2】白洁, 彭建平 , 张康, 等. 基于涡流脉冲热成像方法的稳态闭合疲劳裂纹检测[J].无损检测 , 2018, 40(12):12-16, 86. BAI Jie, PENG Jianping, ZHANG Kang, et al. The detection of steady closed fatigue crack based on ECPT[J]. Nondestructive Testing, 2018, 40(12): 12-16, 86.

【3】LIU Z P, YONG D, ZHU C, et al. Applications of red-green-blue normalization and interframe difference matrix in complex surface by eddy current pulsed thermography[J]. Insight, 2018, 60(12): 670.

【4】PANM C, HEY Z, TIAN G Y, et al. Defect characterization using pulsed eddy current thermography under transmission mode and NDT applications[J]. NDT & E International, 2012,52:28-36.

【5】GAO B, BAI L B, Woo W L, et al. Automatic defect identification of eddy current pulsed thermography using single channel blind source separation[J]. IEEE Transactions on Instrumentation and Measurement, 2014, 63(4): 913-922.

【6】Naderi M, Khonsari M M. On the characterization of thermal-conductivity degradation during torsional fatigue[J]. International Journal of Thermophysics, 2011, 32(3): 693-703.

【7】Morrison A Q, Case E D, Ren F, et al. The elastic modulus, biaxial fracture strength, electrical and thermal transport properties of thermally fatigued hot pressed LAST and LASTT thermoelectric materials[J]. Materials Chemistry and Physics, 2012, 134(2): 973-987.

【8】YIN A J, GAO B, TIAN G Y, et al. Physical interpretation and separation of eddy current pulsed thermography[J]. Journal of Applied Physics, 2013, 064101-1-064101-8. 117-123.

【9】王恒, 周易文 , 季云, 等. 结合狄利克雷过程和连续隐马尔科夫模型 的滚动轴承性能退化评估 [J].吉林大学学报 : 工学版, 2019, 49(1): 113(6): 117-123王恒, 周易文 , 季云, 等. 结合狄利克雷过程和连续隐马尔科夫模型 的滚动轴承性能退化评估 [J].吉林大学学报 : 工学版, 2019, 49(1): 113(6): 117-123

【10】Malik Z K, Hussain A, WU J. An online generalized eigen value version of Laplacian eigen maps for visual big data[J]. Neurocomputing, 2016, 173: 127-136.
WANG Heng, ZHOU Yiwen, JI Yun, et al. Performance degradation assessment of rolling bearing based on Dirichlet process and continuous Markovmodel[J]. Journal of Jilin University: Engineering and Technology Edition, 2019, 49(1): 117-123.

【11】JIA P, YIN J S, HUANGX S, et al . Incremental Laplacian eigenmaps by preserving adjacent information between data points[J]. Pattern Recognition Letters, 2009, 30(16): 1457-1463.

【12】张西宁, 雷威, 杨雨薇 , 等.采用自适应基因粒子群算法优化隐马尔科夫模型的方法及应用[J].西安交通大学学报 , 2018, 52(8): 1-8.

【13】王绵斌, 安磊, 李芬花 , 等. 基于遗传算法的耦合隐马尔科夫模型的故障诊断方法[J].机械设计与制造, 2018(12): 195-198.
WANG Mianbin, AN Lei, LI Fenhua, et al. Fault diagnosis method based on GA-CHM [J]. Machinery Design & Manufacture, 2018(12):

【14】ZHANG Xining, LEI Wei, YANG Yuwei, et al. Adaptive genetic particle swarm algorithm for optimization hidden Markov models with application[J]. Journal of Xi’an Jiaotong University, 2018, 52(8): 1-8.

引用该论文

YIN Aijun,YAO Wenjie. The Evaluation Method and Application of Hidden Markov in Eddy Current Thermal Imaging[J]. Infrared Technology, 2019, 41(12): 1141-1145

尹爱军,姚文杰. 涡流热成像隐马尔科夫评估方法及应用[J]. 红外技术, 2019, 41(12): 1141-1145

您的浏览器不支持PDF插件,请使用最新的(Chrome/Fire Fox等)浏览器.或者您还可以点击此处下载该论文PDF