1.School of Mechatronics and Vehicle Engineering, Chongqing Jiaotong University, Chongqing 400074, China;
2.Key Laboratory of Magnetic Suspension Technology and Maglev Vehicle, Ministry of Education, Chengdu 610031, China;
3.The State Key Laboratory of Mechanical Transmission, Chongqing University, Chongqing 400044, China;
4.Chongqing Industrial Big Data Innovation Center Co., Ltd., Chongqing 404100, China;
5.Chongqing Changjiang Bearing Co., Ltd., Chongqing 401336, China
摘要针对滚动轴承退化性能难以评估、寿命状态难以识别的问题，提出一种基于相空间欧式距离相关性(Phase euclidean distance cross-correlation, PEDCC)指标和多通道卷积长短时记忆网络(Multichannel convolutional neural long short term memory network, MCRNN)的状态识别新方法。首先将正常轴承样本信号进行相空间重构，计算样本内重构后相邻数据之间的欧式距离，并将样本内的所有欧氏距离构成距离向量；然后利用互相关函数计算其余样本距离向量与正常样本距离向量之间的相关性，并将其作为轴承退化指标；最后利用所建立的PEDCC退化指标对轴承状态进行划分，将其输入到MCRNN网络中进行退化状态识别。其中MCRNN网络在不同通道中分别采取了不同卷积核，不同激活函数，以便于提取轴承振动信号的多尺度特征。通过轴承全寿命数据集对所提退化指标及网络模型的实用性进行验证，实验证明所提出的方法能更精确的实现轴承的退化状态识别。
Abstract：Aiming at the difficult problem in evaluating the degradation performance of rolling bearing and identifying bearing running state, a new method based on Phase euclidean distance cross-correlation (PEDCC) indicator and Multichannel convolutional neural long short term memory network (MCRNN) are proposed. Firstly, the normal bearing sample signals were reconstructed in phase space, the Euclidean distances between the reconstructed adjacent data in the sample were calculated, and all the Euclidean distances in the sample were composed into distance vectors. Then, the correlation between the distance vector of the remaining sample and the distance vector of normal sample was calculated by using the cross-correlation function, which was used as the bearing degradation indicator. Finally, the established PEDCC degradation indicator is conducive to the classification of bearing states, which are input into the MCRNN network for degradation state identification. The MCRNN network adopts different convolution kernels and different activation functions in different channels, so as to extract multi-scale features of bearing vibration signals. The practicability of the proposed degradation indicator and network model was verified by the bearing whole-life data set, and the experimental results show that the proposed method can realize the degradation status identification of bearings more accurately.
Key words: rolling bearing; degradation indicator; phase euclidean distance cross-correlation; multichannel convolutional neural long short term memory network; state recognition
肖家丰1，董绍江1,2，汤宝平3，潘雪娇1，胡小林4，赵兴新5. 基于PEDCC性能退化指标及MCRNN的滚动轴承寿命状态识别方法[J]. 振动与冲击, 2022, 41(24): 176-183.
XIAO Jiafeng1,DONG Shaojiang1,2,TANG Baoping3,PAN Xuejiao1,HU Xiaolin4,ZHAO Xingxin5. Rolling bearing life state recognition based on a PEDCC performance degradation indicator and MCRNN. JOURNAL OF VIBRATION AND SHOCK, 2022, 41(24): 176-183.
[1] WU D Y, WANG J W, WANG H, et al. An Automatic Bearing Fault Diagnosis Method Based on Characteristics Frequency Ratio[J]. Sensors, 2020, 20(5).
[2] DONG Shaojiang, WU Wenliang, HE Kun, et al. Rolling bearing performance degradation assessment based on improved convolutional neural network with anti-interference[J]. Measurement, 2020, 151.
[3] WANG R, JIN J H, HU X, et al. Bearing performance degradation assessment based on topological representation and hidden Markov model[J]. Journal of Vibration and Control, 2021, 27(13-14): 1617-1628.
[4] LV M Y, ZHANG C G, GUO A B, et al. A new performance degradation evaluation method integrating PCA, PSR and KELM[J]. IEEE ACCESS, 2021, 9: 6188-6200.
[5] XU X Q, ZHAO M, LIN J, et al. Envelope harmonic-to-noise ratio for periodic impulses detection and its application to bearing diagnosis [J]. Measurement, 2016, 91: 385-397.
[6] ZHOU J M, GUO H J, ZHANG L, et al. Bearing Performance Degradation Assessment Using Lifting Wavelet Packet Symbolic Entropy and SVDD[J]. Shock and Vibration, 2016, 2016.
[7] CHEGINI S N, MANGILI M J H, BAGHERI A. New fault diagnosis approaches for detecting the bearing slight degradation[J], MECCANICA, 2020, 55(1): 261-286.
[8] CHE C C, WANG H W, NI X M, et al. Domain Adaptive Deep Belief Network for Rolling Bearing Fault Diagnosis[J]. Computers & Industrial Engineering, 2020, 143.
[9] 董绍江，裴雪武，吴文亮等. 基于多层降噪技术及改进卷积神经网络的滚动轴承故障诊断方法[J]. 机械工程学报，2021, 57(01): 148-156.
DONG S J, PEI X W, WU W L et al. Rolling Bearing Fault Diagnosis Method Based on Multilayer Noise Reduction Technology and Improved Convolutional Neural Network [J]. Journal of Mechanical Engineering, 2021, 57(01): 148-156.
[10] 李恒，张氢，秦仙蓉，等. 基于短时傅里叶变换和卷积神经网络的轴承故障诊断方法[J]. 振动与冲击，2018, 37(19): 124-131
LI Heng, ZHANG Qing, QIN Xianrong, et al. Fault diagnosis method for rolling bearings based on short-time Fourier transform and convolution neural network[J]. Journal of Vibration and Shock, 2018, 37(19): 124-131.
[11] XU Z, ZHONG L, ZHANG A G. Phase Space Reconstruction Based Conceptor Network for Time Series Prediction[J]. IEEE-ACCESS, 2019, 7: 163172-163179.
[12] KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet classification with deep convolutional neural networks[J]. Communications of the ACM, 2017, 60(6): 84-90.
[13] GREFF K, SRIVASTAVA R K, KOUTNIK J, et al. LSTM: A Search Space Odyssey[J]. IEEE Transactions on Neural Networks and Learning Systems, 2017, 28(10): 2222-2232.
[14] SHE D M, JIA M P. Wear indicator construction of rolling bearings based on multi-channel deep convolutional neural network with exponentially decaying learning rate[J]. Measurement,2019, 135: 368-375.
[15] 雷亚国，韩天宇，王彪等. XJTU-SY 滚动轴承加速 寿命试验数据集解读[J].机械工程学报, 2019, 55(16): 1-6.
LEI Yaguo, HAN Tianyu, WANG Biao, et al. XJTU-SY rolling bearing accelerated life test data set interpretation [J]. Chinese Journal of Mechanical Engineering, 2019, 55 (16): 1-6.
[16] ZHANG W, L C, P G, et al. A deep convolutional neural network with new training methods for bearing fault diagnosis under noisy environment and different working load[J]. Mechanical Systems and Signal Processing, 2018,100:439-453.
