Remaining useful life prediction of aircraft engine based on degradation pattern learning

Zhao, Zeqi; Liang, Bin; Wang, Xueqian; Lu, Weining

doi:10.1016/j.ress.2017.02.007

Cited by 195 publications

(68 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The final step is to label the life of each engine for the next model training. A large number of cases indicate that the label value has a significant impact on prognostic performance [35]. For this dataset, a piece-wise linear degradation model has proven to be appropriate and effective [36].…”

Section: Figure 4: Mon Corr and Sc Of 21 Sensors In Fd001mentioning

confidence: 99%

Ensemble Recurrent Neural Network-Based Residual Useful Life Prognostics of Aircraft Engines

Wu¹,

Hu²,

Cheng³

et al. 2019

Structural Durability &Amp; Health Monitoring

View full text Add to dashboard Cite

Residual useful life (RUL) prediction is a key issue for improving efficiency of aircraft engines and reducing their maintenance cost. Owing to various failure mechanism and operating environment, the application of classical models in RUL prediction of aircraft engines is fairly difficult. In this study, a novel RUL prognostics method based on using ensemble recurrent neural network to process massive sensor data is proposed. First of all, sensor data obtained from the aircraft engines are preprocessed to eliminate singular values, reduce random fluctuation and preserve degradation trend of the raw sensor data. Secondly, three kinds of recurrent neural networks (RNN), including ordinary RNN, long shortterm memory (LSTM), and gated recurrent unit (GRU), are individually constructed. Thirdly, ensemble learning mechanism is designed to merge the above RNNs for producing a more accurate RUL prediction. The effectiveness of the proposed method is validated using two characteristically different turbofan engine datasets. Experimental results show a competitive performance of the proposed method in comparison with typical methods reported in literatures.

show abstract

Section: Figure 4: Mon Corr and Sc Of 21 Sensors In Fd001mentioning

confidence: 99%

Ensemble Recurrent Neural Network-Based Residual Useful Life Prognostics of Aircraft Engines

Wu¹,

Hu²,

Cheng³

et al. 2019

Structural Durability &Amp; Health Monitoring

View full text Add to dashboard Cite

show abstract

“…Lifetime prediction is crucial in manufacturing, automotive, aerospace, and many other industries for supporting an effective decision‐making process . For a highly reliable product, accelerated degradation tests are frequently employed in this process to reduce the test time and to infer the product's lifetime under its normal operating condition …”

Section: Introductionmentioning

confidence: 99%

Ensemble learning for predicting degradation under time‐varying environment

Wang

Lu²,

Wang

et al. 2020

Quality & Reliability Eng

View full text Add to dashboard Cite

Product lifetime prediction is challenging when the product is subject to a time-varying operational environment. Most of the existing studies use some functions to explicitly specify the relationship between degradation parameters and environmental conditions so as to reveal how the degradation process evolves over time. However, in many applications, the assumptions needed for establishing these functions cannot be validated in engineering practice or they cannot accurately model the entire underlying degradation mechanism. In contrast to previous work, the focus of our study is placed on product degradation prognosis by implementing an ensemble learning method. This method combines the stochastic process modeling approach and the machine learning approach, taking advantage of these approaches to gain a more accurate and stable degradation prediction. The proposed method is demonstrated by some simulation examples and by a case study of lithium-ion battery accelerated degradation test. Both the simulation study and the real case verify the superiority of the proposed method. The case study indicates that the ensemble learning method can further help to effectively manage the energy storage and energy distribution of battery packs. K E Y W O R D Sdegradation process, ensemble model, Li-ion battery, machine learning, stochastic process

show abstract

“…For the case when one single sensor cannot show all the features of the degradation system, Liu et al considered data fusion and then presented a novel indicator by integrating multi‐sensor observations under a constraint of the signal‐to‐noise ratio metric. Zhao et al exploited both neural network and learned network as a pattern learning‐based method, for the purpose of discovering the relation between the multivariate time series degradation data and the RULs. Song and Liu combined multiple sensor observations to constitute a composite HI for each component and used the quantile regression to derive the fusion coefficient.…”

Section: Introductionmentioning

confidence: 99%

Remaining useful life prediction for multivariable stochastic degradation systems with non‐Markovian diffusion processes

Zhou

Chen

et al. 2020

Quality & Reliability Eng

View full text Add to dashboard Cite

Multivariable stochastic degradation system (MSDS) is quite common in indus-tries such as blast furnace ironmaking, vehicle transportation, and aerospace manufacturing. Large-scale complex equipments may be affected by multiple factors, resulting in not just a single deteriorating performance characteristic. It is difficult to handle unknown failure structures of practical systems by using traditional univariate degradation modeling methods. A novel health index (HI) is constructed to quantitatively analyze the health state for the overall system. Considering the interaction between internal reactions and external environments, the fractional Brownian motion (FBM), a typical non-Markovian diffusion process, is added for the purpose of reflecting stochastic uncertainties and memory effects. Based on the wavelet estimators and the maximum likelihood estimation (MLE) algorithm, multi-sensor observations of degradation variables are analyzed simultaneously to identify model parameters. A closed-form distribution of system-level remaining useful life (RUL) is obtained with a mild two-layer approximation. Relevant case studies are then handled that adequately demonstrate the effectiveness and the practical utility of the proposed method.

show abstract

Remaining useful life prediction of aircraft engine based on degradation pattern learning

Cited by 195 publications

References 36 publications

Ensemble Recurrent Neural Network-Based Residual Useful Life Prognostics of Aircraft Engines

Ensemble Recurrent Neural Network-Based Residual Useful Life Prognostics of Aircraft Engines

Ensemble learning for predicting degradation under time‐varying environment

Remaining useful life prediction for multivariable stochastic degradation systems with non‐Markovian diffusion processes

Contact Info

Product

Resources

About