Multiscale Dynamic Weighted and Multilevel Residual Convolution Autoencoder Based Rotating Mechanical Signals Denoising

doi:10.3901/JME.2024.18.053

Abstract

Abstract: Vibration signals are used to monitor the running state of rotating machinery, but always suffer from a lot of noise. In order to eliminate noise interference as much as possible, a new noise-learning based neural network was proposed, namely, a multiscale dynamic weight and multilevel residual convolution autoencoder (MDW-MRSCAE) based signals denoising model. In specifically, through the combination of deep convolutional autoencoder and residual module, the autoencoder network can encode and decode the noise components, and replace the traditional learning clean signal by learning the noise characteristics. In order to solve the problem of disappearing gradient in network learning and improve the convergence speed, a multilevel residual structure is constructed. In particular, the multiscale feature extraction network layer is added to enhance the feature extraction ability of the network. In order to reflect the global contribution of different convolution kernels, a dynamic weighted network layer is constructed to adaptively adjust the global weight of each convolution kernel, and further improve the denoising ability of the network. The effectiveness of network denoising was verified by typical analog signals and actual bearing signals with different faults. The results show that the network model is superior to the latest published models of the same kind, and the high frequency noise of the signal is suppressed obviously.

Key words: noise learning, multiscale convolution kernel, dynamic weight, multilevel residual, signal denoising

CLC Number:

TH133
TH165

DU Wenliao, YANG Lingkai, WANG Hongchao, GONG Xiaoyun, ZHAO Feng, LI Chuan. Multiscale Dynamic Weighted and Multilevel Residual Convolution Autoencoder Based Rotating Mechanical Signals Denoising[J]. Journal of Mechanical Engineering, 2024, 60(18): 53-63.

References

[1] 叶壮，余建波. 基于多通道加权卷积神经网络的齿轮箱振动信号特征提取[J]. 机械工程学报，2021，57(1)：110-120. YE Zhuang，YU Jianbo. Feature extraction of gearbox vibration signals based on multi-channels weighted convolutional neural network[J]. Journal of Mechanical Engineering，2021，57(1)：110-120.
[2] 邵海东，肖一鸣，颜深. 仿真数据驱动的改进无监督域适应轴承故障诊断[J].机械工程学报，2023，59(3)：76-85. SHAO Haidong，XIAO Yiming，YAN Shen. Simulation data-driven enhanced unsupervised domain adaptation for bearing fault diagnosis[J]. Journal of Mechanical Engineering，2023，59(3)：76-85.
[3] LIU Dongdong，CUI Lingli，CHENG Weidong，et al. Rolling bearing fault severity recognition via data mining integrated with convolutional neural network[J]. IEEE Sensors Journal，2022，22(6)：5768-5777.
[4] ZHAO Dezun，WANG Tianyang，CHU Fulei. Deep convolutional neural network based planet bearing fault classification[J]. Computers in Industry，2019，107：59-66.
[5] ZHANG Kai，ZUO Wangmeng，CHEN Yunjin，et al. Beyond a Gaussian denoiser：Residual learning of deep CNN for image denoising[J]. IEEE Transactions on Image Processing，2016，26(7)：3142-3155.
[6] CHIANG H T，HSIEH Y Y，FU S W，et al. Noise reduction in ECG signals using fully convolutional denoising autoencoders[J]. IEEE Access，2019，7：60806-60813.
[7] DOUMANIDIS C C，ANAGNOSTOU C，ARVANITI E S，et al. RNNoise-Ex：Hybrid speech enhancement system based on RNN and spectral features[J]. arXiv preprint arXiv：2105，2021，11813.
[8] MIAO Mengqi，LIU Changhui，YU Jianbo. Adaptive densely connected convolutional auto-encoder-based feature learning of gearbox vibration signals[J]. IEEE Transactions on Instrumentation and Measurement，2021，70：1-11.
[9] LIU Xingchen，ZHOU Qicai，ZHAO Jiong，et al. Fault diagnosis of rotating machinery under noisy environment conditions based on a 1-D convolutional autoencoder and 1-D convolutional neural network[J]. Sensors，2019，19(4)：1-20.
[10] HE Kaiming，ZHANG Xiangyu，REN Shaoqing，et al. Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)，2016：770-778.
[11] FAN Gao，LI Jun，HAO Hong. Vibration signal denoising for structural health monitoring by residual convolutional neural networks[J]. Measurement，2020，157：107651.
[12] 周兴康，余建波. 基于深度一维残差卷积自编码网络的齿轮箱故障诊断[J]. 机械工程学报，2020，56(7)：96-108. ZHOU Xingkang，YU Jianbo. Gearbox fault diagnosis based on one-dimension residual convolutional auto-encoder[J]. Journal of Mechanical Engineering，2020，56(7)：96-108.
[13] JIANG Guoqian，HE Haibo，YAN Jun，et al. Multiscale convolutional neural networks for fault diagnosis of wind turbine gearbox[J]. IEEE Transactions on Industrial Electronics，2019，66(4)：3196-3207.
[14] 张成帆，江泽鹏，曹伟，等. 一维多尺度卷积神经网络及其在滚动轴承故障诊断中的应用[J]. 机械科学与技术，2022，41(1)：120-126. ZHANG Chengfan，JIANG Zepeng，CAO Wei，et al. One-dimensional multiscale convolution neural network and its application in rolling bearing fault diagnosis[J]. Mechanical Science and Technology for Aerospace Engineering，2022，41(1)：120-126.
[15] 邓飞跃，丁浩，郝如江. 基于多尺度特征融合残差神经网络的旋转机械故障诊断[J]. 振动与冲击，2021，40(24)：22-28，35. DENG Feiyue，DING Hao，HAO Rujiang. Fault diagnosis of rotating machinery based on residual neural network with multi-scale feature fusion[J]，Journal of Vibration and Shock，2021，40(24)：22-28，35.
[16] JIA Yachao，LI Guolong，DONG Xin，et al. A novel denoising method for vibration signal of hob spindle based on EEMD and grey theory[J]. Measurement，2021，169：108490.
[17] SADOOGHI M S，KHADEM S E. A new performance evaluation scheme for jet engine vibration signal denoising[J]. Mechanical Systems and Signal Processing，2016，76-77：201-212.
[18] DONG Xin，LI Guolong，JIA Yachao，et al. Non-iterative denoising algorithm for mechanical vibration signal using spectral graph wavelet transform and detrended fluctuation analysis[J]. Mechanical Systems and Signal Processing，2021，149：107202.
[19] NAVEED K，AKHTAR M T，SIDDIQUI M F，et al. A statistical approach to signal denoising based on data-driven multiscale representation[J]. Digital Signal Processing，2021，108：102896.
[20] 唐贵基，王晓龙. IVMD融合奇异值差分谱的滚动轴承早期故障诊断[J]. 振动·测试与诊断，2016，36(4)：700-707，810. TANG Guiji，WANG Xiaolong. An incipient fault diagnosis method for rolling bearings based on improved variational mode decomposition and singular value difference spectrum[J]. Journal of Vibration，Measurement & Diagnosis，2016，36(4)：700-707，810.