可解释机械智能诊断技术的研究进展与挑战

doi:10.3901/JME.2023.20.215

摘要/Abstract

摘要： 基于深度神经网络的机械智能诊断方法近年来得到快速发展，各式各样的模型方法相继涌现。然而，此类方法的性能彰显主要是在实验室环境，实际工业场景鲜有应用。究其原因主要在于模型内部非线性变换复杂，特征提取机理不清晰，进而引发使用者对决策的信任问题。特别是对于一些关键设备，倘若不能提前知晓模型得到诊断结论的原由，贸然采取措施会隐藏相当大的风险。近年来，可解释机械智能故障诊断研究越来越多地被关注，一些学者已开始对该领域进行探索，并初见成效。为促进深化研究，推动该领域发展，首先对深度神经网络可解释性研究的各种范式进行分类和讨论，然后详述了可解释性机械智能诊断的当前发展，最后讨论了现有的挑战以及未来的研究方向。

关键词: 深度神经网络, 可解释性, 机械装备, 智能故障诊断

Abstract: The intelligent diagnosis method of machinery based on deep neural networks has developed rapidly in recent years, and various model methods have emerged one after another. However, the performance of such methods is mainly in the laboratory environment, and there are few applications in actual industrial scenarios. The main reason is that the nonlinear transformation inside the model is quite complex, and the feature extraction mechanism is difficult to understand, leading to the user's untrust in decision-making. Especially for some key equipment, if the reason why the model gets the diagnosis conclusion cannot be known in advance, taking measures rashly will hide considerable risks. In light of this, more and more attention has been paid to the interpretability of intelligent fault diagnosis recently, and some scholars have reached some preliminary conclusions. In order to deepen research and promote the development of the field, various paradigms of interpretability of deep neural networks are categorized and discussed, then, current developments in interpretable intelligent fault diagnosis of machinery are detailed, and finally, existing challenges and future research directions are discussed and summarized.

Key words: deep neural networks, interpretability, mechanical equipment, intelligent fault diagnosis

中图分类号:

TG156

林京, 焦金阳. 可解释机械智能诊断技术的研究进展与挑战[J]. 机械工程学报, 2023, 59(20): 215-224.

LIN Jing, JIAO Jinyang. Research Progress and Challenges of Interpretable Mechanical Intelligent Diagnosis[J]. Journal of Mechanical Engineering, 2023, 59(20): 215-224.

导出引用管理器 EndNote|Reference Manager|ProCite|BibTeX|RefWorks

链接本文: http://www.cjmenet.com.cn/CN/10.3901/JME.2023.20.215

http://www.cjmenet.com.cn/CN/Y2023/V59/I20/215

参考文献

[1] 周济. 智能制造——“中国制造2025”的主攻方向[J]. 中国机械工程，2015，26(17):2273-2284. ZHOU Ji. Intelligent manufacturing:Main direction of “Made in China 2025”[J]. China Mechanical Engineering，2015，26，26(17):2273-2284.
[2] 林京. 机器信息学:机械产品智能化的学科支撑[J]. 机械工程学报，2021，57(2):11-20. LIN Jing. Machinery informatics:A fundamental discipline to intelligent machinery[J]. Journal of Mechanical Engineering，2021，57(2):11-20.
[3] 屈梁生. 人工神经网络与机械工程中的智能化问题[J]. 中国机械工程，1997(2):1-4. QU Liangsheng. Artifical neural networks and intelligent issues in mechanical engineering[J]. China Mechanical Engineering，1997(2):1-4.
[4] JIAO J，ZHAO M，LIN J，et al. A comprehensive review on convolutional neural network in machine fault diagnosis[J]. Neurocomputing，2020，417:36-63.
[5] KARNIADAKIS G E，KEVERKIDIS I G，LU L，et al. Physics-informed machine learning[J]. Nature Reviews Physics，2021，3(6):422-440.
[6] MILLER T. Explanation in artificial intelligence:Insights from the social sciences[J]. Artificial Intelligence，2019，267:1-38.
[7] BEUCLER T，PRITCHARD M，RASP S，et al. Enforcing analytic constraints in neural networks emulating physical systems[J]. Physical Review Letters，2021，126(9):98302.
[8] RAISSI M，PERDIKARIS P，KARNIADAKIS G E. Physics-informed neural networks:A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations[J]. Journal of Computational Physics，2019，378:686-707.
[9] WANG N，ZHANG D，CHANG H，et al. Deep learning of subsurface flow via theory-guided neural network[J]. Journal of Hydrology，2020，584:124700.
[10] CHEN Y，HUANG D，ZHANG D，et al. Theory-guided hard constraint projection (HCP):A knowledge-based data-driven scientific machine learning method[J]. Journal of Computational Physics，2021，445:110624.
[11] RAVANELLI M，BENJIO Y. Interpretable convolutional filters with sincnet[J]. arXiv:1811.09725，2018.
[12] TOMEN N，GEMERT Van J C. Spectral leakage and rethinking the kernel size in cnns[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2021:5138-5147.
[13] AMEY J L，KEELEY J，CHOUDHURY T，et al. Neural network interpretation using descrambler groups[J]. Proceedings of the National Academy of Sciences，2021，118(5):e2016917118.
[14] YANG Y，SOATTO S. Fda:Fourier domain adaptation for semantic segmentation[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2020:4085-4095.
[15] PARK S，KOH Y，JEOH H，et al. Enhancing the interpretability of transcription factor binding site prediction using attention mechanism[J]. Scientific Reports，2020，10(1):1-10.
[16] MONGA V，LI Y，ELDAR Y C. Algorithm unrolling: Interpretable，efficient deep learning for signal and image processing[J]. IEEE Signal Processing Magazine，2021，38(2):18-44.
[17] SULAM J，ABERDAM A，BECK A，et al. On multi-layer basis pursuit，efficient algorithms and convolutional neural networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence，2019，42(8):1968-1980.
[18] ZHANG J，GHANEM B. ISTA-Net:Interpretable optimization-inspired deep network for image compressive sensing[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2018:1828-1837.
[19] LI Y，TOFIGHI M，GENG J，et al. Efficient and interpretable deep blind image deblurring via algorithm unrolling[J]. IEEE Transactions on Computational Imaging，2020，6:666-681.
[20] ZEILER M D，FERGUS R. Visualizing and understanding convolutional networks [C]//Computer Vision–ECCV 2014:13th European Conference，Zurich，Switzerland，September 6-12，2014，Proceedings，Part I 13. Springer International Publishing，2014:818-833.
[21] SHRIKUMAR A，GREENSIDE P，KUNDAJE A. Learning important features through propagating activation differences[C]//International Conference on Machine Learning, PMLR，2017:3145-3153.
[22] SMIKOV D，THORAT N，KIM B，et al. Smoothgrad: Removing noise by adding noise[J]. arXiv:1706.03825，2017.
[23] SRINIVAS S，FLEURET F. Full-gradient representation for neural network visualization[C]//Conference on Neural Information Processing Systems，Vancouver，December 12, 2019, 2020:4101-4110.
[24] ZHOU B，KHOLSA A，LAPEDRIZA A，et al. Learning deep features for discriminative localization [C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2016:2921-2929.
[25] SELVARAJU R R，COGSWELL M，DAS A，et al. Grad-cam: Visual explanations from deep networks via gradient-based localization[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2017:618-626.
[26] WANG H，WANG Z，DU M，et al. Score-CAM: Score-weighted visual explanations for convolutional neural networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2020:24-25.
[27] CHATTOPADHAY A，SARKAR A，HOWLADER P，et al. Grad-CAM++: Generalized Gradient-Based Visual Explanations for Deep Convolutional Networks [C]//2018 IEEE winter Conference on Applications of Computer Vision (WACV). IEEE，2018:839-847.
[28] ZHAO G，ZHOU B，WANG K，et al. Respond-cam: Analyzing deep models for 3d imaging data by visualizations[C]//Medical Image Computing and Computer Assisted Intervention–MICCAI 2018: 21st International Conference，Granada，Spain，September 16-20，2018，Proceedings，Part I. Springer International Publishing，2018:485-492.
[29] SAURABH D，RAMASWAMY H G，Ablation-cam: Visual explanations for deep convolutional network via gradient-free localization[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2020:983-991.
[30] RIBEIRO M T，SINGH S，GUESTRIN C. “Why should I trust you?”Explaining the predictions of any classifier[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining，2016:1135-1144.
[31] BRAMHALL S，HORN H，TIEU M，et al. Qlime-a quadratic local interpretable model-agnostic explanation approach[J]. SMU Data Science Review，2020，3(1):4.
[32] AHERN I，NOACK A，GUZMAN-NATERAS L，et al. NormLime: A new feature importance metric for explaining deep neural networks[J]. arXiv:1909.04200，2019.
[33] TAO G，MA S，LIU Y，et al. Attacks meet interpretability: Attribute-steered detection of adversarial samples[J]. Advances in Neural Information Processing Systems，2018，31:7717-7728.
[34] ABDUKHAMIDOV E，ABUHAMAD M，WOO S S，et al. Interpretations cannot be trusted:Stealthy and effective adversarial perturbations against interpretable deep learning[J]. arXiv:2211.15926，2022.
[35] MOLNAR C. Interpretable machine learning[M]. New York:Lulu. Com.，2020.
[36] FANG B，HU J，YANG C，et al. A blind deconvolution algorithm based on backward automatic differentiation and its application to rolling bearing fault diagnosis[J]. Measurement Science and Technology，2021，33(2):25009.
[37] MIAO Y，LI C，SHI H，et al. Deep network-based maximum correlated kurtosis deconvolution: A novel deep deconvolution for bearing fault diagnosis[J]. Mechanical Systems and Signal Processing，2023，189:110110.
[38] SHANG Z，ZHAO Z，YAN R. Denoising fault-aware wavelet network: A signal processing informed neural network for fault diagnosis[J]. Chinese Journal of Mechanical Engineering，2023，36(1):9.
[39] LI T，ZHAO Z，SUN C，et al. WaveletKernelNet: An interpretable deep neural network for industrial intelligent diagnosis[J]. IEEE Transactions on Systems，Man，and Cybernetics:Systems，2021，52(4):2302-2312.
[40] LI T J，SUN C，LI S，et al. Explainable graph wavelet denoising network for intelligent fault diagnosis[J]. IEEE Transactions on Neural Networks and Learning Systems，2022，12:1-14.
[41] CHEN Y，RAO M，FENG K，et al. Physics-Informed LSTM hyperparameters selection for gearbox fault detection[J]. Mechanical Systems and Signal Processing，2022，171:108907.
[42] LI Y，ZHOU Z，SUN C，et al. Variational attention-based interpretable transformer network for rotary machine fault diagnosis[J]. IEEE Transactions on Neural Networks and Learning Systems，2022，9:1-14.
[43] WANG D，CHEN Y，SHEN C，et al. Fully interpretable neural network for locating resonance frequency bands for machine condition monitoring[J]. Mechanical Systems and Signal Processing，2022，168:108673.
[44] ZHANG H，LIN J，HUA J，et al. Interpretable convolutional sparse coding method of Lamb waves for damage identification and localization[J]. Structural Health Monitoring，2022，21(4):1790-1804.
[45] AN B，WANG S，ZHAO Z，et al. Interpretable neural network via algorithm unrolling for mechanical fault diagnosis[J]. IEEE Transactions on Instrumentation and Measurement，2022，71:1-11.
[46] ZHAO Z，LI T，AN B，et al. Model-driven deep unrolling: Towards interpretable deep learning against noise attacks for intelligent fault diagnosis[J]. ISA Transactions，2022，129:644-662.
[47] YANG H，LI X，ZHANG W. Interpretability of deep convolutional neural networks on rolling bearing fault diagnosis[J]. Measurement Science and Technology，2022，33(5):55005.
[48] CHENG W，WANG S，LIU Y，et al. A novel planetary gearbox fault diagnosis method for nuclear circulating water pump with class imbalance and data distribution shift[J]. IEEE Transactions on Instrumentation and Measurement，2023，72:1-13.
[49] YU S，WANG M，PANG S，et al. Intelligent fault diagnosis and visual interpretability of rotating machinery based on residual neural network[J]. Measurement，2022，196:111228.
[50] KIM M S，YUN J P，PARK P. An explainable convolutional neural network for fault diagnosis in linear motion guide[J]. IEEE Transactions on Industrial Informatics，2020，17(6):4036-4045.
[51] SUN H，CAO X，WANG C，et al. An interpretable anti-noise network for rolling bearing fault diagnosis based on FSWT[J]. Measurement，2022，190:110698.
[52] TANG J，ZHENG G，WEI C，et al. Signal-transformer: A robust and interpretable method for rotating machinery intelligent fault diagnosis under variable operating conditions[J]. IEEE Transactions on Instrumentation and Measurement，2022，71:1-11.
[53] LU F，TONG Q，FENG Z，et al. Explainable 1DCNN with demodulated frequency features method for fault diagnosis of rolling bearing under time-varying speed conditions[J]. Measurement Science and Technology，2022，33(9):95022.
[54] ZHANG H，LIN J，HUA J，et al. Attention-based interpretable prototypical network towards small-sample damage identification using ultrasonic guided waves[J]. Mechanical Systems and Signal Processing，2023，188:109990.
[55] TANG R，ZHANG S，WU W，et al. Explainable deep learning based ultrasonic guided wave pipe crack identification method[J]. Measurement，2023，206:112277.
[56] GWAK M，KIM M S，YUN J P，et al. Robust and explainable fault diagnosis with power-perturbation-based decision boundary analysis of deep learning models[J]. IEEE Transactions on Industrial Informatics，2022，19(5):6982-6992.