Statistical Alignment-Based Metagated Recurrent Unit for Cross-Domain Machinery Degradation Trend Prognostics Using Limited Data

Nowadays, machinery degradation trend prognostics have made significant progress with sufficient and related data. Limited data may weaken the generalizations of artificial intelligence-based prognostics, which could worsen under the domain shift phenomenon. To overcome the dual challenges and further abstract the prior degraded knowledge from historical data, a novel prediction architecture entitled as statistical alignment-based metagated recurrent unit, profoundly integrating deep learning, knowledge transfer, and metalearning, is proposed in this study and simultaneously features domain adaptation ability and desired generalizations under limited data. To be specific, a designed high-order measure is first developed to align training and testing data distribution. Then, the well-known encoder–decoder framework is employed as the base model, and subtask & cross-subtask learning architectures are developed to distill sensitive prognostics agents for easily generalizing to new domains using limited data. Finally, FEMTO-ST bearing data sets and industrial data sets collected from petrochemical pumps are conducted to prove their effectiveness.

[1]  Paul Maropoulos,et al.  A novel method for accurately monitoring and predicting tool wear under varying cutting conditions based on meta-learning , 2019, CIRP Annals.

[2]  Brigitte Chebel-Morello,et al.  PRONOSTIA : An experimental platform for bearings accelerated degradation tests. , 2012 .

[3]  Liang Gao,et al.  A New Two-Level Hierarchical Diagnosis Network Based on Convolutional Neural Network , 2020, IEEE Transactions on Instrumentation and Measurement.

[4]  Jun Zhu,et al.  A new data-driven transferable remaining useful life prediction approach for bearing under different working conditions , 2020 .

[5]  Yoshua Bengio,et al.  Learning Phrase Representations using RNN Encoder–Decoder for Statistical Machine Translation , 2014, EMNLP.

[6]  Minping Jia,et al.  Health condition identification for rolling bearing using a multi-domain indicator-based optimized stacked denoising autoencoder , 2020 .

[7]  Yi Qin,et al.  Gated Dual Attention Unit Neural Networks for Remaining Useful Life Prediction of Rolling Bearings , 2021, IEEE Transactions on Industrial Informatics.

[8]  Nishchal K. Verma,et al.  Intelligent Condition-Based Monitoring of Rotary Machines With Few Samples , 2020, IEEE Sensors Journal.

[9]  Tao Zhang,et al.  Deep Model Based Domain Adaptation for Fault Diagnosis , 2017, IEEE Transactions on Industrial Electronics.

[10]  Zhibin Zhao,et al.  Few-shot transfer learning for intelligent fault diagnosis of machine , 2020 .

[11]  Yaguo Lei,et al.  A New Method Based on Stochastic Process Models for Machine Remaining Useful Life Prediction , 2016, IEEE Transactions on Instrumentation and Measurement.

[12]  Qiang Yang,et al.  A Survey on Transfer Learning , 2010, IEEE Transactions on Knowledge and Data Engineering.

[13]  Balázs Kégl,et al.  Intrinsic Dimension Estimation Using Packing Numbers , 2002, NIPS.

[14]  Ricardo Vilalta,et al.  A Perspective View and Survey of Meta-Learning , 2002, Artificial Intelligence Review.

[15]  Robert X. Gao,et al.  An integrated fault diagnosis and prognosis approach for predictive maintenance of wind turbine bearing with limited samples , 2020 .

[16]  Noura Dridi,et al.  Akaike and Bayesian Information Criteria for Hidden Markov Models , 2019, IEEE Signal Processing Letters.

[17]  Xiang Li,et al.  Remaining useful life estimation in prognostics using deep convolution neural networks , 2018, Reliab. Eng. Syst. Saf..

[18]  Sergey Levine,et al.  Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks , 2017, ICML.

[19]  Peng Ding,et al.  A dynamic structure-adaptive symbolic approach for slewing bearings’ life prediction under variable working conditions , 2020 .

[20]  Xuefeng Chen,et al.  Conditional Adversarial Domain Adaptation With Discrimination Embedding for Locomotive Fault Diagnosis , 2021, IEEE Transactions on Instrumentation and Measurement.

[21]  Xu Li,et al.  Data alignments in machinery remaining useful life prediction using deep adversarial neural networks , 2020, Knowl. Based Syst..

[22]  Yi Chai,et al.  Macroscopic–Microscopic Attention in LSTM Networks Based on Fusion Features for Gear Remaining Life Prediction , 2020, IEEE Transactions on Industrial Electronics.

[23]  Anoop Cherian,et al.  Jensen-Bregman LogDet Divergence with Application to Efficient Similarity Search for Covariance Matrices , 2013, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[24]  Wei Zhang,et al.  Intelligent rotating machinery fault diagnosis based on deep learning using data augmentation , 2018, Journal of Intelligent Manufacturing.

[25]  Mohd Salman Leong,et al.  Gearbox Fault Diagnosis Using a Deep Learning Model With Limited Data Sample , 2020, IEEE Transactions on Industrial Informatics.

[26]  Xiang Li,et al.  Cross-Domain Fault Diagnosis of Rolling Element Bearings Using Deep Generative Neural Networks , 2019, IEEE Transactions on Industrial Electronics.

[27]  Wentao Mao,et al.  Predicting Remaining Useful Life of Rolling Bearings Based on Deep Feature Representation and Transfer Learning , 2020, IEEE Transactions on Instrumentation and Measurement.

[28]  Jie Chen,et al.  Degradation evaluation of slewing bearing using HMM and improved GRU , 2019, Measurement.

[29]  Wei Zhang,et al.  A robust intelligent fault diagnosis method for rolling element bearings based on deep distance metric learning , 2018, Neurocomputing.

[30]  Liang Zhang,et al.  Fault Diagnosis of Bearing in Wind Turbine Gearbox Under Actual Operating Conditions Driven by Limited Data With Noise Labels , 2021, IEEE Transactions on Instrumentation and Measurement.

[31]  Mehrtash Tafazzoli Harandi,et al.  Learning an Invariant Hilbert Space for Domain Adaptation , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).