Turbofan engine is a critical aircraft component with complex structure and high-reliability requirements.
Effectively predicting the remaining useful life (RUL) of turbofan engines has essential significance for
developing maintenance strategies and reducing maintenance costs. Considering the characteristics of large sample
size and high dimension of monitoring data, a hybrid health condition prediction model integrating the advantages
of autoencoder and bidirectional long short-term memory (BLSTM) is proposed to improve the prediction
accuracy of RUL. Autoencoder is used as a feature extractor to compress condition monitoring data. BLSTM
is designed to capture the bidirectional long-range dependencies of features. A hybrid deep learning prediction
model of RUL is constructed. This model has been tested on a benchmark dataset. The results demonstrate
that this autoencoder-BLSTM hybrid model has a better prediction accuracy than the existing methods, such as
multi-layer perceptron (MLP), support vector regression (SVR), convolutional neural network (CNN) and long
short-term memory (LSTM). The proposed model can provide strong support for the health management and
maintenance strategy development of turbofan engines.
SONG Ya (宋亚), SHI Guo (石郭), CHEN Leyi (陈乐懿), HUANG Xinpei (黄鑫沛), XIA Tangbin (夏唐斌)
. Remaining Useful Life Prediction of Turbofan Engine Using Hybrid Model Based on Autoencoder and Bidirectional Long Short-Term Memory[J]. Journal of Shanghai Jiaotong University(Science), 2018
, 23(Sup. 1)
: 85
-94
.
DOI: 10.1007/s12204-018-2027-5
[1] FENG D L, XIAO M Q, LIU Y X, et al. A kernelprincipal component analysis-based degradation modeland remaining useful life estimation for the turbofanengine [J]. Advances in Mechanical Engineering, 2016,8(5): 1-13.
[2] XIA T B, DONG Y F, XIAO L, et al. Recent advancesin prognostics and health management for advancedmanufacturing paradigms [J]. Reliability Engineeringand System Safety, 2018, 178: 255-268.
[3] NIETO P J G, GARC′IA-GONZALO E, LASHERASF S, et al. Hybrid PSO-SVM-based method for forecastingof the remaining useful life for aircraft enginesand evaluation of its reliability [J]. Reliability Engineeringand System Safety, 2015, 138: 219-231.
[4] NG S S Y, XING Y J, TSUI K L. A naive Bayes modelfor robust remaining useful life prediction of lithiumionbattery [J]. Applied Energy, 2014, 118: 114-123.
[5] DEUTSCH J, HE D. Using deep learning-based approachto predict remaining useful life of rotating components[J]. IEEE Transactions on Systems, Man, andCybernetics: Systems, 2018, 48(1): 11-20.
[6] BABU G S, ZHAO P L, LI X L. Deep convolutionalneural network based regression approach for estimationof remaining useful life [C]//International Conferenceon Database Systems for Advanced Applications.Dallas, USA: LNCS, 2016: 214-228.
[7] ZHANG Y Z, XIONG R, HE H W, et al. Long shorttermmemory recurrent neural network for remaininguseful life prediction of lithium-ion batteries [J]. IEEETransactions on Vehicular Technology, 2018, 67(7):5695-5705.
[8] ZHENG S, RISTOVSKI K, FARAHAT A, et al. Longshort-term memory network for remaining useful lifeestimation [C]//International Conference on Prognosticsand Health Management. Florida, USA: IEEE,2017: 88-95.
[9] ZEYER A, DOETSCH P, VOIGTLAENDER P, et al.A comprehensive study of deep bidirectional LSTMRNNs for acoustic modeling in speech recognition[C]//International Conference on Acoustics, Speechand Signal Processing. New Orleans, USA: IEEE, 2017:2462-2466.
[10] ALTHELAYA K A, EL-ALFY E S M, MOHAMMEDS. Evaluation of bidirectional LSTM for short- andlong-term stock market prediction [C]//InternationalConference on Information and Communication Systems.Irbid, Jordan: IEEE, 2018: 151-156.
[11] JOLLIFFE I T, CADIMA J. Principal componentanalysis: A review and recent developments [J].Philosophical Transactions of the Royal Society A:Mathematical, Physical & Engineering Sciences, 2016,374(2065): 20150202.
[12] ZHAO M B, ZHANG Z, CHOW T W, S et al. A generalsoft label based linear discriminant analysis forsemi-supervised dimensionality reduction [J]. NeuralNetworks, 2014, 55: 83-97.
[13] HINTON G E, SALAKHUTDINOV R R. Reducingthe dimensionality of data with neural networks [J].Science, 2006, 313(5786): 504-507.
[14] WANG Y S, YAO H X, ZHAO S C, et al. Auto-encoderbased dimensionality reduction [J]. Neurocomputing,2016, 184: 232-242.
[15] SAXENA A, GOEBEL K, SIMON D, et al. Damagepropagation modeling for aircraft engine run-to-failuresimulation [C]//International Conference on Prognosticsand Health Management. Denver, Colorado, USA:IEEE, 2008: 1-9.
[16] WU Y T, YUAN M, DONG S P, et al. Remaininguseful life estimation of engineered systems usingvanilla LSTM neural networks [J]. Neurocomputing,2018, 275: 167-179.
[17] ARCOS-GARC′IA ′ A, ′ALVAREZ-GARC′IA J A,SORIA-MORILLO L M. Deep neural network for trafficsign recognition systems: An analysis of spatialtransformers and stochastic optimisation methods [J].Neural Networks, 2018, 99: 158-165.
[18] HEIMES F O. Recurrent neural networks for remaininguseful life estimation [C]//International Conferenceon Prognostics and Health Management. Denver,Colorado, USA: IEEE, 2008: 59-64.
