面向滚动轴承全生命周期故障诊断的GA-OIHF Elman神经网络算法

doi:10.16183/j.cnki.jsjtu.2020.157

摘要/Abstract

摘要：

针对高背景噪声下滚动轴承全生命周期(轻度退化、中度退化、重度退化)故障诊断需求,提出GA-OIHF (Genetic Algorithm-Output Input Hidden Feedback) Elman神经网络模型,实现退化故障的精准诊断.利用集合经验模态分解对振动信号进行有效降噪与故障特征提取.设计OIHF Elman神经网络,并在Elman神经网络结构的基础上,同时增加输出层对隐含层与输入层的反馈,进一步提高其对滚动轴承全生命周期数据的处理能力.然后,通过结合遗传算法构建一种新的GA-OIHF Elman神经网络模型,该模型综合了遗传算法的全局寻优与OIHF Elman神经网络的局部寻优能力,从而实现对滚动轴承全生命周期的精确故障诊断.实验结果表明,所提出的GA-OIHF Elman方法不仅对于滚动轴承全生命周期故障具有准确的诊断效果,而且保证了诊断模型对于不同故障(不同故障部件与不同故障时期)的诊断稳定性.

关键词: 滚动轴承, 遗传算法, Elman神经网络, 集合经验模态分解, 全生命周期

Abstract:

For the fault diagnosis needs of the full life cycle (light degradation, moderate degradation, and severe degradation) of rolling bearing under the environment of high background noise, a genetic algorithm-output input hidden feedback (GA-OIHF ) Elman neural network model is proposed to achieve precise diagnosis of the degradation faults of rolling bearing. Ensemble empirical mode decomposition (EEMD) is selected to effectively reduce the noise and extract fault features of the vibration signal. An OIHF Elman neural network is designed by increasing the feedbacks from the output layer to the hidden layer and the input layer based on the Elman neural network, thus further improves its ability to process full life cycle data of rolling bearing. Then, a novel GA-OIHF Elman neural network model is developed by combining the genetic algorithm (GA) and the OIHF Elman neural network. The novel GA-OIHF Elman neural network model combines the global optimization of GA and the local optimization ability of the OIHF Elman neural network to achieve an accurate fault diagnosis of the entire life cycle of rolling bearing. The experimental results show that the GA-OIHF Elman algorithm model can not only accurately diagnose the fault in the full life cycle of rolling bearing, but also ensure the stability of the diagnosis model for different faults including different fault components and stages.

Key words: rolling bearing, genetic algorithm, Elman neural network, ensemble empirical mode decomposition (EEMD), full life cycle

中图分类号:

TH17

卓鹏程, 严瑾, 郑美妹, 夏唐斌, 奚立峰. 面向滚动轴承全生命周期故障诊断的GA-OIHF Elman神经网络算法[J]. 上海交通大学学报, 2021, 55(10): 1255-1262.

ZHUO Pengcheng, YAN Jin, ZHENG Meimei, XIA Tangbin, XI Lifeng. GA-OIHF Elman Neural Network Algorithm for Fault Diagnosis of Full Life Cycle of Rolling Bearing[J]. Journal of Shanghai Jiao Tong University, 2021, 55(10): 1255-1262.

图/表 10

图1

图2

图3

图4

图5

图6

表1

SVM与4种神经网络在测试集上的平均诊断结果

诊断模型	$MSE ¯$	STD
SVM	0.0804	0.0978
Elman	0.0609	0.0868
OHF Elman	0.0204	0.0582
OIF Elman	0.0201	0.0616
OIHF Elman	0.0131	0.0470

表1

表2

3种进化算法下的OIHF Elman神经网络模型的诊断结果

诊断模型	$MSE ¯$	STD
OIHF Elman	0.0131	0.0470
GA-OIHF Elman	0.0100	0.0403
PSO-OIHF Elman	0.0114	0.0438
MEA-OIHF Elman	0.0115	0.0423

表2

表3

基于GA-OIHF Elman神经网络模型的全生命周期故障诊断MSE

故障状态与故障部件	神经网络类型	$MSE ¯$
正常	OIHF Elman	3.21×10^-14
正常	GA-OIHF Elman	0.0087
轻度退化	OIHF Elman	0.0022
轻度退化	GA-OIHF Elman	0.0043
中度退化	OIHF Elman	0.0316
中度退化	GA-OIHF Elman	0.0186
重度退化	OIHF Elman	0.0069
重度退化	GA-OIHF Elman	0.0036
滚动体	OIHF Elman	0.0061
滚动体	GA-OIHF Elman	0.0079
内圈	OIHF Elman	0.0113
内圈	GA-OIHF Elman	0.0039
外圈	OIHF Elman	0.0270
外圈	GA-OIHF Elman	0.0164

表3

表4

GA-OIHF Elman神经网络模型在3种部件的不同退化阶段诊断MSE

退化故障阶段	故障部件	$MSE ¯$
轻度退化	内圈	1.36×10^-15
	外圈	2.97×10^-9
	滚动体	0.0101
中度退化	内圈	1.35×10^-178
	外圈	0.0319
	滚动体	0.0111
重度退化	内圈	0.0101
	外圈	2.05×10^-13
	滚动体	1.24×10^-31

表4

参考文献 19

[1]	HUANG W T, SUN H J, LUO J N, et al. Periodic feature oriented adapted dictionary free OMP for rolling element bearing incipient fault diagnosis[J]. Mechanical Systems and Signal Processing, 2019, 126:137-160. doi: 10.1016/j.ymssp.2019.02.023 URL
[2]	XIA T B, DONG Y F, XIAO L, et al. Recent advances in prognostics and health management for advanced manufacturing paradigms[J]. Reliability Engineering & System Safety, 2018, 178:255-268. doi: 10.1016/j.ress.2018.06.021 URL
[3]	XIA T B, SONG Y, ZHENG Y, et al. An ensemble framework based on convolutional bi-directional LSTM with multiple time windows for remaining useful life estimation[J]. Computers in Industry, 2020, 115:103182. doi: 10.1016/j.compind.2019.103182 URL
[4]	SAARI J, STRÖMBERGSSON D, LUNDBERG J, et al. Detection and identification of windmill bearing faults using a one-class support vector machine (SVM)[J]. Measurement, 2019, 137:287-301. doi: 10.1016/j.measurement.2019.01.020 URL
[5]	陈玉昇, 杨燕华, 林萌, 等. 基于深度学习神经网络的核电厂故障诊断技术[J]. 上海交通大学学报, 2018, 52(Sup.1):58-61.
	CHEN Yusheng, YANG Yanhua, LIN Meng, et al. Fault diagnosis technology of nuclear power plant based on deep learning neural network[J]. Journal of Shanghai Jiao Tong University, 2018, 52(Sup.1):58-61.
[6]	BEN ALI J, FNAIECH N, SAIDI L, et al. Application of empirical mode decomposition and artificial neural network for automatic bearing fault diagnosis based on vibration signals[J]. Applied Acoustics, 2015, 89:16-27. doi: 10.1016/j.apacoust.2014.08.016 URL
[7]	LI L, HUANG Y X, TAO J F, et al. Featured temporal segmentation method and AdaBoost-BP detector for internal leakage evaluation of a hydraulic cylinder[J]. Measurement, 2018, 130:279-289. doi: 10.1016/j.measurement.2018.08.029 URL
[8]	汤宝平, 习建民, 李锋. 基于Elman神经网络的旋转机械故障诊断[J]. 计算机集成制造系统, 2010, 16(10):2148-2152.
	TANG Baoping, XI Jianmin, LI Feng. Fault diagnosis for rotating machinery based on Elman neural network[J]. Computer Integrated Manufacturing Systems, 2010, 16(10):2148-2152.
[9]	SHI X H, LIANG Y C, LEE H P, et al. Improved Elman networks and applications for controlling ultrasonic motors[J]. Applied Artificial Intelligence, 2004, 18(7):603-629. doi: 10.1080/08839510490483279 URL
[10]	ZHENG K, YANG D W, ZHANG B, et al. A group sparse representation method in frequency domain with adaptive parameters optimization of detecting incipient rolling bearing fault[J]. Journal of Sound and Vibration, 2019, 462:114931. doi: 10.1016/j.jsv.2019.114931 URL
[11]	LEI Y G, QIAO Z J, XU X F, et al. An underdamped stochastic resonance method with stable-state matching for incipient fault diagnosis of rolling element bearings[J]. Mechanical Systems and Signal Processing, 2017, 94:148-164. doi: 10.1016/j.ymssp.2017.02.041 URL
[12]	LI H, LIU T, WU X, et al. Application of EEMD and improved frequency band entropy in bearing fault feature extraction[J]. ISA Transactions, 2019, 88:170-185. doi: 10.1016/j.isatra.2018.12.002 URL
[13]	HUANG N E, SHEN Z, LONG S R, et al. The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis[J]. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 1998, 454(1971):903-995. doi: 10.1098/rspa.1998.0193 URL
[14]	WANG Z W, ZHANG Q H, XIONG J B, et al. Fault diagnosis of a rolling bearing using wavelet packet denoising and random forests[J]. IEEE Sensors Journal, 2017, 17(17):5581-5588. doi: 10.1109/JSEN.2017.2726011 URL
[15]	AMIRAT Y, BENBOUZID M E H, WANG T, et al. EEMD-based notch filter for induction machine bearing faults detection[J]. Applied Acoustics, 2018, 133:202-209. doi: 10.1016/j.apacoust.2017.12.030 URL
[16]	刘涛, 杜世昌, 黄德林, 等. 基于改进的集合经验模态方法振动信号分解[J]. 上海交通大学学报, 2016, 50(9):1452-1459.
	LIU Tao, DU Shichang, HUANG Delin, et al. Vibration signal decomposition based on an improved ensemble empirical mode decomposition method[J]. Journal of Shanghai Jiao Tong University, 2016, 50(9):1452-1459.
[17]	FU Q, JING B, HE P J, et al. Fault feature selection and diagnosis of rolling bearings based on EEMD and optimized Elman_AdaBoost algorithm[J]. IEEE Sensors Journal, 2018, 18(12):5024-5034. doi: 10.1109/JSEN.2018.2830109 URL
[18]	GUO H X, LI Y J, LI Y N, et al. BPSO-Adaboost-KNN ensemble learning algorithm for multi-class imbalanced data classification[J]. Engineering Applications of Artificial Intelligence, 2016, 49:176-193. doi: 10.1016/j.engappai.2015.09.011 URL
[19]	LIU H, TIAN H Q, LIANG X F, et al. New wind speed forecasting approaches using fast ensemble empirical model decomposition, genetic algorithm, Mind Evolutionary Algorithm and Artificial Neural Networks[J]. Renewable Energy, 2015, 83:1066-1075. doi: 10.1016/j.renene.2015.06.004 URL