基于改进鲸鱼算法优化LSTM的滚动轴承故障诊断

摘要
图/表
参考文献(26)
相关文章 (15)

全文: PDF (1687 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要适应性动量(Adam)估计优化器易使深度长短时记忆神经网络(long short-term memory，LSTM)陷入局部极小值，导致故障诊断精度过低；鲸鱼算法(whale optimization algorithm，WOA)的寻优区域过大，导致寻优效率过低。针对上述两问题，将WOA进行改进(improved whale optimization algorithm，IWOA)，并优化LSTM，提出IWOA-LSTM新方法。所提方法将WOA赋予动量驱动功能，继承了LSTM中的Adam优化器动量项，可优化细胞权值的搜索区域，进而提高权值寻优效率；将其与Adam优化器联合优化更新权值矩阵，以跳出局部最小值，提高故障诊断精度。此外，还系统地分析了学习效率和迭代次数对IWOA-LSTM的诊断精度影响，实现高效的故障诊断分析。通过分析实测滚动轴承内圈、外圈和滚动体三种故障可知，IWOA-LSTM的故障诊断效率分别较浅层BP神经网络(BPNN)、深度卷积神经网络(convolutional neural network ，CNN)、深度门限循环单元神经网络(gated recurrent unit，GRU)、LSTM、WOA优化的LSTM(WOA-LSTM)高出了47.60%,38.06%,37.62%,26.82%,22.71%，且实现高达97%的诊断精度。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	郑直1
	2
	张华钦1
	潘月3

关键词 ：滚动轴承, 深度学习, 鲸鱼算法(WOA), 长短时记忆(LSTM)

Abstract：The adaptive moment (Adam) estimation optimizer tends to make the long short-term memory (LSTM) neural network fall into local minimum value, which leads to excessive low accuracy of rolling bearing fault diagnosis. The searching area of the whale optimization algorithm (WOA) is too large, which leads to excessive low optimization efficiency. Here, aiming at the above two problems, WOA was improved, and LSTM was optimized to propose a new IWOA-LSTM method. It was shown that the proposed method endows WOA with momentum driven function and inherits the momentum term of Adam optimizer in LSTM to optimize the search area of cell weight value and improve the efficiency of weight value optimization; then, WOA is combined with Adam optimizer to update the weight value matrix to jump out of the local minimum value and improve the accuracy of fault diagnosis; in addition, effects of learning efficiency and iteration times on the diagnosis accuracy of IWOA-LSTM is systematically analyzed to realize highly effective fault diagnosis analysis. Three kinds of faults of inner ring, outer ring and rolling element of actual bearing data measured were analyzed, the results showed that the fault diagnosis efficiency of IWOA-LSTM is 47.60%, 38.06%, 37.62%, 26.82% and 22.71% higher than those of shallow BP neural network (BPNN), deep convolution neural network (CNN), deep gated recurrent unit neural network (GRU), LSTM and WOA-LSTM, respectively; the diagnosis accuracy is up to 97%.

Key words： roll bearing deep learning whale optimization algorithm (WOA) long short-term memory (LSTM)

收稿日期: 2019-12-30 出版日期: 2021-04-15

引用本文:

郑直1,2，张华钦1，潘月3. 基于改进鲸鱼算法优化LSTM的滚动轴承故障诊断[J]. 振动与冲击, 2021, 40(7): 274-280.
ZHENG Zhi1,2, ZHANG Huaqin1, PAN Yue3. Rolling bearing fault diagnosis based on IWOA-LSTM. JOURNAL OF VIBRATION AND SHOCK, 2021, 40(7): 274-280.

链接本文:

http://jvs.sjtu.edu.cn/CN/ 或 http://jvs.sjtu.edu.cn/CN/Y2021/V40/I7/274

［1］沈长青, 谢伟达, 朱忠奎, 等. 基于EEMD和改进的形态滤波方法的轴承故障诊断研究［J］. 振动与冲击, 2013, 32(2):39-43.
SHEN Changqing, XIE Weida, ZHU Zhongkui, et al. Rolling element bearing fault diagnosis based on EEMD and improved morphological filtering method［J］. Journal of Vibration and Shock,2013,32(2):39-43.
［2］LU S L, GUO J, HE Q B, et al. A novel contactless angular resampling method for motor bearing fault diagnosis under variable speed［J］. IEEE Transactions on Instrumentation & Measurement, 2016, 65(11):2538-2550.
［3］ZHANG X N, HAN Q K, PENG Z K, et al. Stability analysis of a rotor-bearing system with time-varying bearing stiffness due to finite number of balls and unbalanced force［J］. Journal of Sound & Vibration, 2013, 332(25):6768-6784.
［4］李志农, 张芬, 肖尧先. 基于CVA-ICA的机械故障源动态盲分离方法［J］. 机械工程学报, 2015, 51(12):24-29.
LI Zhinong, ZHANG Fen, XIAO Yaoxian. Dynamic blind separation of mechanical fault sources based on canonical variate analysis and independent component analysis［J］. Journal of Mechanical Engineering, 2015, 51(12):24-29.
［5］ZHENG J D, PAN H P, CHENG J S. Rolling bearing fault detection and diagnosis based on composite multiscale fuzzy entropy and ensemble support vector machines［J］. Mechanical Systems & Signal Processing, 2017, 85:746-759.
［6］李巍华, 翁胜龙, 张绍辉. 一种萤火虫神经网络及在轴承故障诊断中的应用［J］. 机械工程学报, 2015(7): 99-106.
LI Weihua, WENG Shenglong, ZHANG Shaohui.A firefly neural network and its application in bearing fault diagnosis［J］. Journal of Mechanical Engineering, 2015(7):99-106.
［7］程军圣, 史美丽, 杨宇. 基于LMD与神经网络的滚动轴承故障诊断方法［J］. 振动与冲击, 2010, 29(8): 141-144.
CHENG Junsheng, SHI Meili, YANG Yu.Roller bearing fault diagnosis method based on LMD and neural network［J］. Journal of Vibration and Shock,2010,29(8):141-144.
［8］冯志鹏, 宋希庚, 薛冬新. 基于粗糙集与神经网络集成的内燃机故障诊断［J］. 内燃机学报, 2003, 21(1):75-80.
FENG Zhipeng, SONG Xigeng, XUE Dongxin.Fault diagnosis based on integration of rough sets and neural networks［J］. Transactions of CSICE, 2003, 21(1):75-80.
［9］HINTON G E, SALAKHUTDINOV R R.Reduing the dimensionality of data with neural networks［J］. Science, 2006, 313(5786): 504-507.
［10］XU X F, LEI Y G, LI Z D.An incorrect data detection method for big data cleaning of machinery condition monitoring［J］. IEEE Transactions on Industrial Electronics, 2020, 67(3):2326-2336.
［11］雷亚国, 贾峰, 周昕, 等. 基于深度学习理论的机械装备大数据健康监测方法［J］. 机械工程学报, 2015, 51(21): 49-56.
LEI Yaguo, JIA Feng, ZHOU Xin, et al. A deep learning-based method for machinery health monitoring with big data［J］. Journal of Mechanical Engineering, 2015, 51(21):49-56.
［12］杨宇,张娜,程军圣. 全参数动态学习深度信念网络在滚动轴承寿命预测中的应用［J］. 振动与冲击, 2019, 38(10): 199-205.
YANG Yu, ZHANG Na, CHENG Junsheng. Global parameters dynamic learning deep belief networks and its application in rolling bearing life prediction［J］. Journal of Vibration and Shock, 2019, 38(10):199-205.
［13］WANG X, FENG L, WANG J X, et al. Quantum weighted gated recurrent unit neural network and its application in performance degradation trend prediction of rotating machinery［J］. Neurocomputing, 2018, 313: 85-95.
［14］GUO L, LI N P, JIA F, et al. A recurrent neural network based health indicator for remaining useful life prediction of bearings［J］. Neurocomputing, 2017, 240:98-109.
［15］SHAO H D, JIANG H K, ZHANG H Z, et al. Rolling bearing fault feature learning using improved convolutional deep belief network with compressed sensing［J］. Mechanical Systems & Signal Processing, 2018, 100:743-765.
［16］PENG D D, LIU Z L, WANG H, et al. A novel deeper one-dimensional CNN with residual learning for fault diagnosis of wheelset bearings in high-speed trains［J］. IEEE Access, 2018,7:10278-10293.
［17］SHE D, JIA M Q. 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.
［18］GRAVES A. Long short-term memory［M］. Berlin: Springer, 2012．
［19］王鑫, 吴际, 刘超, 等. 基于LSTM循环神经网络的故障时间序列预测［J］. 北京航空航天大学学报, 2018, 44(4):772-784.
WANG Xin, WU Ji, LIU Chao, et al.Exploring LSTM based recurrent neural network forfailure time series prediction［J］. Journal of Beijing University of Aeronautics and Astronautics, 2018, 44(4):772-784.
［20］YU L, QU J L, GAO F, et al. A novel hierarchical algorithm for bearing fault diagnosis based on stacked LSTM［J］. Shock and Vibration, 2019(12):1-10.
［21］ZHANG J J, WANG P, YAN R Q, et al. Long short-term memory for machine remaining life prediction［J］. Journal of Manufacturing Systems, 2018, 48:78-86.
［22］ALJARAH I, FARIS H, MIRJALILI S. Optimizing connection weights in neural networks using the whale optimization algorithm［J］. Soft Computing, 2018, 22(1):1-15.
［23］MIAO Y H, ZHAO M, MAKIS V, et al. Optimal swarm decomposition with whale optimization algorithm for weak feature extraction from multicomponent modulation signal［J］. Mechanical Systems and Signal Processing, 2019, 2:673-691.
［24］赵春华,胡恒星,陈保家, 等.基于深度学习特征提取和WOA-SVM状态识别的轴承故障诊断［J］.振动与冲击,2019,38(10):31-37.
ZHAO Chunhua，HU Hengxing，CHEN Baojia，et al. Bearing fault diagnosis based on the deep learning feature extraction and WOA-SVM state recognition［J］. Journal of Vibration and Shock, 2019, 38(10):31-37.
［25］许瑜飞, 钱锋, 杨明磊, 等. 改进鲸鱼优化算法及其在渣油加氢参数优化的应用［J］. 化工学报, 2018, 69(3):891-899.
XU Yufei, QIAN Feng, YANG Minglei, et al. Improved whale optimization algorithm and its application in optimization of residue hydrogenation parameters［J］. Journal of Chemical Industry and Engineering, 2018, 69(3):891-899.
［26］REDDY K S, PANWAR L, PANI8GRAHI B K, et al. Binary whale optimization algorithm: a new metaheuristic approach for profit-based unit commitment problems in competitive electricity markets［J］.Engineering Optimization, 2019, 51(3):369-389.