基于可解释选择性集成框架的离心风机叶片裂纹损伤检测

doi:10.3901/JME.2024.12.183

机械工程学报 ›› 2024, Vol. 60 ›› Issue (12): 183-193.doi: 10.3901/JME.2024.12.183

• 特邀专栏：可解释可信AI驱动的智能监测与诊断 • 上一篇下一篇

扫码分享

基于可解释选择性集成框架的离心风机叶片裂纹损伤检测

沈君贤, 马天池, 宋狄, 许飞云

东南大学机械工程学院南京 211102

收稿日期:2023-09-05 修回日期:2024-03-09 出版日期:2024-06-20 发布日期:2024-08-23
作者简介:沈君贤,男,1995年出生,博士研究生。主要研究方向为机械故障诊断、结构健康监测。E-mail:230218566@seu.edu.cn;许飞云(通信作者),男,1969年出生,博士,教授,博士研究生导师。主要研究方向为设备状态监控与故障诊断、动态系统建模与辨识。E-mail:fyxu@seu.edu.cn
基金资助:
国家自然科学基金资助项目(51975117)。

Crack Damage Detection of Centrifugal Fan Blades Based on Interpretable Ensemble Selection Framework

SHEN Junxian, MA Tianchi, SONG Di, XU Feiyun

School of Mechanical Engineering, Southeast University, Nanjing 211102

Received:2023-09-05 Revised:2024-03-09 Online:2024-06-20 Published:2024-08-23

摘要/Abstract

摘要： 接触式测量的振动信号是最为常用的旋转设备状态监测手段，但在工业现场离心风机叶片断裂前会产生异常的声响，然而其振动信号却依旧平稳。声信号属于一种非接触测量方法，包含强烈的背景噪声激励，使得其数据驱动网络的可信任程度较低。因此，为有效增强声信号驱动模型的可解释性、评估模型决策的可信度，提出一种可解释的选择性集成框架用于叶片裂纹的检测。首先，采用Alexnet提取的深度特征和可解释的辅助统计特征结合注意力权重机制构建一个多视角的基网络。其次，将声信号时频图像构建二阶张量作为网络的输入，对训练好的每个基网络的深度特征通过Grad-CAM确定激活映射面积并结合仿真结果计算网络的可信任度。接着，对于嵌入式的辅助特征构建一种Diversity Pick LIME(DP-LIME)的可解释模块，结合特征权重分布可视化基网络决策逻辑。最终，根据每个基网络的综合可信任度指标对叶片裂纹检测结果进行选择性决策融合。经离心风机实测数据验证，文中提出的可解释框架具有良好的叶片裂纹检测精度，并可有效提高模型可信度。

关键词: 离心风机叶片, 损伤检测, 选择性集成, 可解释模块, 特征注意力

Abstract: Contact-measured vibration signals is the most commonly used means of health monitoring for rotating equipment. However, in the industrial site, the centrifugal fan blade will produce abnormal sounds before breaking, but the vibration signal is still stable. The acoustic signal is a non-contact measurement method and contains strong background noise excitation, making its data-driven network less trustworthy. Therefore, to effectively enhance the interpretability of acoustic signal-driven models and evaluate the credibility of model decisions, an interpretable ensemble selection framework for blade crack detection is proposed. First, a multi-view base network is constructed by adopting Alexnet-extracted depth features and interpretable auxiliary statistical features combined with an attentional weighting mechanism. Secondly, the second-order tensor of the time-frequency image of the acoustic signal is used as input to the network, and the depth features of each trained base network are adopted to determine the activation mapping area by Grad-CAM and to calculate the trustworthiness of the network in conjunction with the simulation results. Next, a Diversity Pick LIME (DP-LIME) interpretable module is constructed for embedded auxiliary features, combined with a feature weight distribution to visualize the decision logic in the base network. Finally, the blade crack detection results are selectively fused for decision-making based on the composite trustworthiness index of each base network. The interpretable framework proposed in this paper has high detection accuracy for blade crack and can effectively improve the credibility of the model, which is verified by the actual measurement data of the centrifugal fan.

Key words: centrifugal fan blades, damage detection, ensemble selection, interpretable modules, feature attention mechanism

中图分类号:

TH165
TH17

沈君贤, 马天池, 宋狄, 许飞云. 基于可解释选择性集成框架的离心风机叶片裂纹损伤检测[J]. 机械工程学报, 2024, 60(12): 183-193.

SHEN Junxian, MA Tianchi, SONG Di, XU Feiyun. Crack Damage Detection of Centrifugal Fan Blades Based on Interpretable Ensemble Selection Framework[J]. Journal of Mechanical Engineering, 2024, 60(12): 183-193.

参考文献

[1] GAI J B，SHEN J X，HU Y F，et al. An integrated method based on hybrid grey wolf optimizer improved variational mode decomposition and deep neural network for fault diagnosis of rolling bearing[J]. Measurement，2020，162(10)：1-12.
[2] 王辉，徐佳文，严如强. 基于多尺度注意力深度强化学习网络的行星齿轮箱智能诊断方法[J]. 机械工程学报，2022，58(11)：133-142. WANG Hui，XU Jiawen，YAN Ruqiang. Multi-scale attention based deep reinforcement learning for intelligent fault diagnosis of planetary gearbox[J]. Journal of Mechanical Engineering，2022，58(11)：133-142.
[3] WANG D，CHEN Y K，SHEN C Q，et al. Fully interpretable neural network for locating resonance frequency bands for machine condition monitoring[J]. Mechanical Systems and Signal Processing，2022，168(11)：1-17.
[4] DING P，JIA M P. Mechatronics equipment performance degradation assessment using limited and unlabeled data[J]. IEEE Transactions on Industrial Informatics，2021，18(4)：2374-2385.
[5] LI W，ZHONG X，SHAO H D，et al. Multi-mode data augmentation and fault diagnosis of rotating machinery using modified ACGAN designed with new framework[J]. Advanced Engineering Informatics，2022，52(4)：1-16.
[6] 周登极，郝佳瑞，黄大文. 机器学习模型可解释性研究及其在PHM中应用现状综述[J]. 系统工程，2022，40(6)：1-10. ZHOU Dengji，HAO Jiarui，HUANG Dawen. A review on interpretable machine learning and its application on PHM[J]. Systems Engineering，2022，40(6)：1-10.
[7] GOODFELLOW I，SHLENS J，SZEGEDY C. Explaining and harnessing adversarial examples[C]// 3rd International Conference on Learning Representations，2015，5(1)：1-11.
[8] SONG D，SHEN J X，MA T C，et al. Two-level fusion of multi-sensor information for compressor blade crack detection based on self-attention mechanism[J]. Structural Health Monitoring，2022，22(3)：1911-1926.
[9] XIE X P，CHEN W D，CHEN B G，et al. Comprehensive fatigue estimation and fault diagnosis based on refined generalized multi-scale entropy method of centrifugal fan blades[J]. Measurement，2020，166(12)：1-17.
[10] NING D Y，SUN C L，GONG Y J，et al. Extraction of fault component from abnormal sound in diesel engines using acoustic signals[J]. Mechanical Systems & Signal Processing，2016，75(6)：544-555.
[11] GLOWACZ A，TADEUSIEWICZ R，LEGUTKO S，et al. Fault diagnosis of angle grinders and electric impact drills using acoustic signals[J]. Applied Acoustics，2021，179(8)：1-14.
[12] WANG R，LIU F K，HOU F T，et al. A non-contact fault diagnosis method for rolling bearings based on acoustic imaging and convolutional neural networks[J]. IEEE Access，2020，8(1)：132761-132774.
[13] KAMAL M S，DEY N，CHOWDHURY L，et al. Explainable AI for glaucoma prediction analysis to understand risk factors in treatment planning[J]. IEEE Transactions on Instrumentation and Measurement，2022，71(5)：1-12.
[14] SAMEK W，MONTAVON G，LAPUSCHKIN S，et al. Explaining deep neural networks and beyond：A review of methods and applications[J]. Proceedings of the IEEE，2021，109(3)：247-278.
[15] AHMED I，JEON G，PICCIALLI F. From artificial intelligence to explainable artificial intelligence in Industry 4.0：A survey on What，How，and Where[J]. IEEE Transactions on Industrial Informatics，2022，18(8)：5031-5042.
[16] 王冉，石如玉，胡升涵，等. 基于声成像与卷积神经网络的轴承故障诊断方法及其可解释性研究[J]. 振动与冲击，2022，41(16)：224-231. WANG Ran，SHI Ruyu，HU Shenghan，et al. An acoustic fault diagnosis method of rolling bearings based on acoustic imaging and convolutional neural network[J]. Journal of Vibration and Shock，2022，41(16)：224-231.
[17] HASAN M，SOHAIB M，KIM J. An explainable AI-based fault diagnosis model for bearings[J]. Sensors，2021，21(12)：1-34.
[18] 宋志坤，徐立成，胡晓依，等. 基于改进型shapelets算法的动车组轴箱轴承故障诊断方法研究[J]. 仪器仪表学报，2021，42(2)：66-74. SONG Zhikun，XU Licheng，HU Xiaoyi，et al. Research on fault diagnosis method of axle box bearing of EMU based on improved shapelets algorithm[J]. Chinese Journal of Scientific Instrument，2021，42(2)：66-74.
[19] WANG Z，ZHAO W，WENHUA D，et al. Data-driven fault diagnosis method based on the conversion of erosion operation signals into images and convolutional neural network[J]. Process Safety and Environmental Protection，2021，149(5)：591-601.
[20] 王松涛，丁骏，姜斌，等. 三维叶片技术对离心风机流场结构的影响[J]. 机械工程学报，2014，50(4)：178-184. WANG Songtao，DING Jun，JIANG Bin，et al. Effect of three-dimensional blade technology on flow behavior in centrifugal fan[J]. Journal of Mechanical Engineering，2014，50(4)：178-184.
[21] 刘卓，汤健，柴天佑，等. 基于多模态特征子集选择性集成建模的磨机负荷参数预测方法[J]. 自动化学报，2021，47(8)：1921-1931. LIU Zhuo，TANG Jian，CHAI Tianyou，et al. Selective ensemble modeling approach for mill load parameter forecasting based on multi-modal feature sub-sets[J]. Acta Automatica Sinica，2021，47(8)：1921-1931.

基于可解释选择性集成框架的离心风机叶片裂纹损伤检测

Crack Damage Detection of Centrifugal Fan Blades Based on Interpretable Ensemble Selection Framework

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 4

编辑推荐

Metrics

本文评价

[1]	张喆, 陈刚, 陈旭. 基于力响应发光材料的结构损伤可视化检测技术研究进展[J]. 机械工程学报, 2023, 59(16): 54-71.
[2]	蔡瑞, 李勇, 刘天浩, 裴翠祥, 陈振茂. 金属小径管损伤电磁超声导波检测的高效混合仿真方法及导波换能器可行性研究[J]. 机械工程学报, 2020, 56(10): 34-41.
[3]	李锐华,徐梓源,胡波,潘玲. 基于双面激励导波的大电机定子绝缘损伤检测[J]. 电气工程学报, 2016, 11(9): 1-7.
[4]	马书义;武湛君;刘科海;王奕首. 管道变形损伤超声导波检测试验研究[J]. , 2013, 49(14): 1-8.