基于双输入深度卷积神经网络的风力机轴承RUL预测

刘杰; 苏宇涵; 陈长征

doi:10.19912/j.0254-0096.tynxb.2022-1283

PDF(4870 KB)

太阳能学报 ›› 2023, Vol. 44 ›› Issue (12) : 238-250. DOI: 10.19912/j.0254-0096.tynxb.2022-1283

基于双输入深度卷积神经网络的风力机轴承RUL预测

刘杰, 苏宇涵, 陈长征

作者信息 +

REMAINING USEFUL LIFE PREDICTION OF WIND TURBINE BEARINGS BASED ON DUAL-INPUT DEEP CONVOLUTIONAL NEURAL NETWORK

Liu Jie, Su Yuhan, Chen Changzheng

Author information +

文章历史 +

摘要

针对轴承剩余寿命预测中常用健康指标泛化性不足的问题,提出一种基于双输入深度卷积神经网络的轴承剩余寿命预测模型。首先使用自适应最大相关峭度解卷积方法处理轴承信号并采用特征融合手段得到信号的时间序列特征;然后,将信号的时频图和时间序列特征同时作为模型的输入,通过已建立的双输入深度卷积神经网络模型来预测轴承健康因子;最后使用门控循环单元网络与健康因子相结合的方法来预测轴承的剩余使用寿命。在公开的西安交通大学公布的XJTU轴承数据集上对所提方法进行验证,并在风力机高速轴轴承历史监测数据上进行应用。试验结果表明：该方法不但显著提升了健康因子的泛化性能,还在预测精度方面有优异表现。

Abstract

To solve the problem of insufficient generalization for common health indicators in the bearing remaining useful life （RUL） prediction, a model of the RUL prediction based on dual-input deep convolutional neural network （DUALCNN） is proposed. Firstly, the bearing signals are processed by an adaptive maximum correlation kurtosis deconvolution （MCKD）, and the time series features can be obtained by fusing features. Then, the time-frequency diagram and time-series features are simultaneously used as the input of the model. The bearing health indicator （HI） can be predicted by establishing the DUALCNN model. Finally, the predicted HI is combined with a gated recurrent unit （GRU） network to predict the RUL of the bearings. The proposed method is validated on the publicly available XJTU bearing datasets and is applied to the historical monitoring data from high-speed shaft bearings in a wind turbine. The experimental results show that the proposed method significantly improves the generalization performance of the HI. Meanwhile, the proposed method has an excellent performance in terms of the accuracy of remaining useful life prediction.

导出引用

刘杰, 苏宇涵, 陈长征. 基于双输入深度卷积神经网络的风力机轴承RUL预测[J]. 太阳能学报. 2023, 44(12): 238-250 https://doi.org/10.19912/j.0254-0096.tynxb.2022-1283

Liu Jie, Su Yuhan, Chen Changzheng. REMAINING USEFUL LIFE PREDICTION OF WIND TURBINE BEARINGS BASED ON DUAL-INPUT DEEP CONVOLUTIONAL NEURAL NETWORK[J]. Acta Energiae Solaris Sinica. 2023, 44(12): 238-250 https://doi.org/10.19912/j.0254-0096.tynxb.2022-1283

中图分类号： TH133.3

参考文献

[1] LIAO H T, TIAN Z G.A framework for predicting the remaining useful life of a single unit under time-varying operating conditions[J]. IIE transactions, 2013, 45(9): 964-980.
[2] 张金豹, 邹天刚, 王敏, 等. 滚动轴承剩余使用寿命预测综述[J]. 机械科学与技术, 2023, 42(1): 1-23.
ZHANG J B, ZOU T G, WANG M, et al.Review on remaining useful life prediction of rolling bearing[J]. Mechanical science and technology for aerospace engineering, 2023, 42(1): 1-23.
[3] 裴迪, 岳建海, 焦静. 基于自相关与能量算子增强的滚动轴承微弱故障特征提取[J]. 振动与冲击, 2021, 40(11): 101-108, 123.
PEI D, YUE J H, JIAO J.Weak fault feature extraction of rolling bearing based on autocorrelation and energy operator enhancement[J]. Journal of vibration and shock, 2021, 40(11): 101-108, 123.
[4] 张龙, 胡俊锋, 熊国良. 基于MED和SK的滚动轴承循环冲击特征增强[J]. 振动测试与诊断, 2017, 37(1): 97-101, 201.
ZHANG L, HU J F, XIONG G L.Cyclic shock enhancement by the combination of minimum entropy deconvolution and spectral kurtosis[J]. Journal of vibration, measurement & diagnosis, 2017, 37(1): 97-101, 201.
[5] MCDONALD G, ZHAO Q, ZUO M.Maximum correlated Kurtosis deconvolution and application on gear tooth chip fault detection[J]. Mechanical systems and signal processing, 2012, 33: 237-255.
[6] 赵洪山, 李浪. 基于最大相关峭度解卷积和变分模态分解的风电机组轴承故障诊断方法[J]. 太阳能学报, 2018, 39(2): 350-358.
ZHAO H S, LI L.Fault diagnosis method of wind turbine bearing based on maximum correlated kurtosis deconvolution and variational mode decomposition[J]. Acta energiae solaris sinica, 2018, 39(2): 350-358.
[7] 陈鑫, 郭瑜, 伍星, 等. 基于MCKD和改进IESFOgram相结合的行星轴承外圈故障诊断[J]. 振动与冲击, 2021, 40(20): 200-206.
CHEN X, GUO Y, WU X, et al.Planet bearing outer-race fault diagnosis based on MCKD and improved IESFOgram[J]. Journal of vibration and shock, 2021, 40(20): 200-206.
[8] LEI Y G, LI N P, GUO L, et al.Machinery health prognostics: a systematic review from data acquisition to RUL prediction[J]. Mechanical systems and signal processing, 2018, 104: 799-836.
[9] SINGLETON R K, STRANGAS E G, AVIYENTE S.Extended Kalman filtering for Remaining-Useful-Life estimation of bearings[J]. IEEE transactions on industrial electronics, 2015, 62(3): 1781-1790.
[10] ZHAO R, YAN R Q, CHEN Z H, et al. Deep learning and its applications to machine health monitoring: a survey[EB/OL]. arXiv:1612.07640. https://arxiv.org/abs/1612.07640, 2016.
[11] LI X Q, JIANG H K, XIONG X, et al.Rolling bearing health prognosis using a modified health index based hierarchical gated recurrent unit network[J]. Mechanism and machine theory, 2019, 133: 229-249.
[12] JAVED K, GOURIVEAU R, ZERHOUNI N, et al.Enabling health monitoring approach based on vibration data for accurate prognostics[J]. IEEE transactions on industrial electronics, 2015, 62(1): 647-656.
[13] WANG Y X, XIANG J, MARKERT R, et al.Spectral kurtosis for fault detection, diagnosis and prognostics of rotating machines: a review with applications[J]. Mechanical systems and signal processing, 2016, 66: 679-698.
[14] LI Y B, WANG X Z, LIU Z B, et al.The entropy algorithm and its variants in the fault diagnosis of rotating machinery: a review[J]. IEEE access, 2018, 6: 66723-66741.
[15] 吕明珠, 苏晓明, 陈长征, 等. 基于PCA-UPF的风力机轴承剩余寿命预测方法[J]. 太阳能学报, 2021, 42(2): 218-224.
LYU M Z, SU X M, CHEN C Z, et al.Prediction approach of remaining useful life for wind turbine bearings based on pca-upf[J]. Acta energiae solaris sinica, 2021, 42(2): 218-224.
[16] 陈昌, 汤宝平, 吕中亮. 基于威布尔分布及最小二乘支持向量机的滚动轴承退化趋势预测[J]. 振动与冲击, 2014, 33(20): 52-56.
CHEN C, TANG B P, LÜ Z L.Degradation trend prediction of rolling bearings based on Weibull distribution and least squares support vector machine[J]. Journal of vibration and shock, 2014, 33(20): 52-56.
[17] 王玉静, 王诗达, 康守强, 等. 基于改进深度森林的滚动轴承剩余寿命预测方法[J]. 中国电机工程学报, 2020, 40(15): 5032-5043.
WANG Y J, WANG S D, KANG S Q, et al.Prediction method of remaining useful life of rolling bearings based on improved GcForest[J]. Proceedings of the CSEE, 2020, 40(15): 5032-5043.
[18] 王冉, 后麒麟, 石如玉, 等. 基于变分模态分解与集成深度模型的锂电池剩余寿命预测方法[J]. 仪器仪表学报, 2021, 42(4): 111-120.
WANG R, HOU Q L, SHI R Y, et al.Remaining useful life prediction method of lithium battery based on variational mode decomposition and integrated deep model[J]. Chinese journal of scientific instrument, 2021, 42(4): 111-120.
[19] 车畅畅, 王华伟, 倪晓梅, 等. 基于1D-CNN和Bi-LSTM的航空发动机剩余寿命预测[J]. 机械工程学报, 2021, 57(14): 304-312.
CHE C C, WANG H W, NI X M, et al.Residual life prediction of aeroengine based on 1D-CNN and Bi-LSTM[J]. Journal of mechanical engineering, 2021, 57(14): 304-312.
[20] 王久健, 杨绍普, 刘永强, 等. 一种基于空间卷积长短时记忆神经网络的轴承剩余寿命预测方法[J]. 机械工程学报, 2021, 57(21): 88-95.
WANG J J, YANG S P, LIU Y Q, et al.A method of bearing remaining useful life estimation based on convolutional long short-term memory neural network[J]. Journal of mechanical engineering, 2021, 57(21): 88-95.
[21] 张俊, 张建群, 钟敏, 等. 基于PSO-VMD-MCKD方法的风机轴承微弱故障诊断[J]. 振动测试与诊断, 2020, 40(2): 287-296, 418.
ZHANG J, ZHANG J Q, ZHONG M, et al.PSO-VMD-MCKD based fault diagnosis for incipient damage in wind turbine rolling bearing[J]. Journal of vibration, measurement & diagnosis, 2020, 40(2): 287-296, 418.
[22] WANG J Y, MO Z L, ZHANG H, et al.A deep learning method for bearing fault diagnosis based on time-frequency image[J]. IEEE access, 2019, 7: 42373-42383.
[23] KRIZHEVSKY A, SUTSKEVER I, HINTON G E.ImageNet classification with deep convolutional neural networks[C]//Proceedings of the 25th International Conference on Neural Information Processing Systems - Volume 1. Lake Tahoe, Nevada, USA, 2012: 1097-1105.
[24] CHO K, VAN MERRIENBOER B, GULCEHRE C, et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation[EB/OL]. arXiv:1406.1078. https://arxiv.org/abs/1406.1078, 2014.
[25] NECTOUX P, GOURIVEAU R, MEDJAHER K, et al.PRONOSTIA: an experimental platform for bearings accelerated degradation tests[C]//IEEE International Conference on Prognostics and Health Management.Denver, USA, 2012.