Transfer Learning with Deep Recurrent Neural Networks for Remaining Useful Life Estimation
Source abstract
Abstract
Prognostics, such as remaining useful life (RUL) prediction, is a crucial task in condition-based maintenance. A major challenge in data-driven prognostics is the difficulty of obtaining a sufficient number of samples of failure progression. However, for traditional machine learning methods and deep neural networks, enough training data is a prerequisite to train good prediction models. In this work, we proposed a transfer learning algorithm based on Bi-directional Long Short-Term Memory (BLSTM) recurrent neural networks for RUL estimation, in which the models can be first trained on different but related datasets and then fine-tuned by the target dataset. Extensive experimental results show that transfer learning can in general improve the prediction models on the dataset with a small number of samples. There is one exception that when transferring from multi-type operating conditions to single operating conditions, transfer learning led to a worse result.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Machine Fault Diagnosis Techniques · Non-Destructive Testing Techniques · Reliability and Maintenance Optimization
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Key Laboratory of Advanced Manufacturing Technology of Ministry of Education, Guizhou University, Guiyang 550025, China
Department of Computer Science and Engineering, University of South Carolina, Columbia, SC 29208, USA
Guizhou Provincial Key Laboratory of Internet Collaborative intelligent manufacturing, Guizhou University, Guiyang 550025, China
School of Mechanical Engineering, Guizhou University, Guiyang 550025, China
Funding
National Natural Science Foundation of China · Project of Ministry of Industry and Information · Science and Technology Project of Guizhou Province
References (37)
Show source-provided references
1. Machinery health prognostics: A systematic review from data acquisition to RUL prediction10.1016/j.ymssp.2017.11.016 ↗
2. Fault diagnosis of rotary machinery components using a stacked denoising autoencoder-based health state identification10.1016/j.sigpro.2016.07.028 ↗
3. Probabilistic physics of failure-based framework for fatigue life prediction of aircraft gas turbine discs under uncertainty10.1016/j.ress.2015.10.002 ↗
4. Data-based techniques focused on modern industry: An overview10.1109/TIE.2014.2308133 ↗
5. Prognostics and health management design for rotary machinery systems—Reviews, methodology and applications10.1016/j.ymssp.2013.06.004 ↗
6. Performance Estimation and Fault Diagnosis Based on Levenberg–Marquardt Algorithm for a Turbofan Engine10.3390/en11010181 ↗
7. Condition based maintenance in railway transportation systems based on big data streaming analysis10.1016/j.procs.2015.07.321 ↗
8. Degradation Modelling for Health Monitoring Systems
9. Khelassi, A., Theilliol, D., Weber, P., and Ponsart, J.C. (2011, January 28–30). Fault-tolerant control design with respect to actuator health degradation: An LMI approach. Proceedings of the 2011 IEEE International Conference on Control Applications (CCA), Denver, CO, USA.10.1109/CCA.2011.6044381 ↗
10. A review on machinery diagnostics and prognostics implementing condition-based maintenance10.1016/j.ymssp.2005.09.012 ↗
11. Eker, O.F., Camci, F., and Jennions, I.K. (2012, January 3–5). Major Challenges in Prognostics: Study on Benchmarking Prognostics Datasets. Proceedings of the First European Conference of the Prognostics and Health Management Society 2012, Dresden, German.10.36001/phme.2012.v1i1.1409 ↗
12. Optimize the signal quality of the composite health index via data fusion for degradation modeling and prognostic analysis10.1109/TASE.2015.2446752 ↗
13. A model-based method for remaining useful life prediction of machinery10.1109/TR.2016.2570568 ↗
14. Feature evaluation for effective bearing prognostics10.1002/qre.1396 ↗
15. Health-state estimation and prognostics in machining processes10.1109/TASE.2009.2038170 ↗
16. Structural health monitoring techniques for aircraft composite structures10.1016/j.paerosci.2010.05.001 ↗
17. Residual life predictions in the absence of prior degradation knowledge10.1109/TR.2008.2011659 ↗
18. A simple state-based prognostic model for railway turnout systems10.1109/TIE.2010.2051399 ↗
19. Deep learning10.1038/nature14539 ↗
20. Li, S., Liu, G., Tang, X., Lu, J., and Hu, J. (2017). An ensemble deep convolutional neural network model with improved ds evidence fusion for bearing fault diagnosis. Sensors, 17.10.3390/s17081729 ↗
21. Li, S., Yao, Y., Hu, J., Liu, G., Yao, X., and Hu, J. (2018). An ensemble stacked convolutional neural network model for environmental event sound recognition. Appl. Sci., 8.10.3390/app8071152 ↗
22. Yao, Y., Wang, H., Li, S., Liu, Z., Gui, G., Dan, Y., and Hu, J. (2018). End-to-end convolutional neural network model for gear fault diagnosis based on sound signals. Appl. Sci., 8.10.3390/app8091584 ↗
23. Babu, G.S., Zhao, P., and Li, X.L. (2016). Deep Convolutional Neural Network Based Regression Approach for Estimation of Remaining Useful Life. International Conference on Database Systems for Advanced Applications, Springer.10.1007/978-3-319-32025-0_14 ↗
24. Zheng, S., Ristovski, K., Farahat, A., and Gupta, C. (2017, January 19–21). Long short-term memory network for remaining useful life estimation. Proceedings of the 2017 IEEE International Conference on IEEE Prognostics and Health Management (ICPHM), Allen, TX, USA.10.1109/ICPHM.2017.7998311 ↗
25. Zhao, R., Yan, R., Wang, J., and Mao, K. (2017). Learning to monitor machine health with convolutional Bi-directional LSTM networks. Sensors (Switzerland), 17.10.3390/s17020273 ↗
26. Remaining useful life estimation of engineered systems using vanilla LSTM neural networks10.1016/j.neucom.2017.05.063 ↗
27. Li, Z., Wu, D., Hu, C., and Terpenny, J. (2018). An ensemble learning-based prognostic approach with degradation-dependent weights for remaining useful life prediction. Reliab. Eng. Syst. Saf.10.1115/DETC2017-68315 ↗
28. A survey on transfer learning10.1109/TKDE.2009.191 ↗
29. Weiss, K., Khoshgoftaar, T.M., and Wang, D.D. (2016). A Survey of Transfer Learning. J. Big Data.10.1186/s40537-016-0043-6 ↗
30. Meng, Z., Chen, Z., Mazalov, V., Li, J., and Gong, Y. (2017, January 16–20). Unsupervised adaptation with domain separation networks for robust speech recognition. Proceedings of the 2017 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU), Okinawa, Japan.10.1109/ASRU.2017.8268938 ↗
31. Singh, K.K., Divvala, S., Farhadi, A., and Lee, Y.J. (2018). DOCK: Detecting Objects by Transferring Common-Sense Knowledge. European Conference on Computer Vision, Springer.10.1007/978-3-030-01261-8_30 ↗
32. Cao, H., Bernard, S., Heutte, L., and Sabourin, R. (2018). Improve the performance of transfer learning without fine-tuning using dissimilarity-based multi-view learning for breast cancer histology images. International Conference Image Analysis and Recognition, Springer.10.1007/978-3-319-93000-8_88 ↗
33. Performance Benchmarking and Analysis of Prognostic Methods for CMAPSS Datasets
34. Saxena, A., and Goebel, K. (2018, November 20). Turbofan Engine Degradation Simulation Data Set. NASA Ames Prognostics Data Repository, Available online: https://ti.arc.nasa.gov/tech/dash/groups/pcoe/prognostic-data-repository/.
35. LONG SHORT-TERM MEMORY10.1162/neco.1997.9.8.1735 ↗
36. Framewise phoneme classification with bidirectional LSTM and other neural network architectures10.1016/j.neunet.2005.06.042 ↗
37. Heimes, F. (2008, January 6–9). Recurrent Neural Networks for Remaining Useful Life Estimation. Proceedings of the International Conference on 2008 Prognostics and Health Management, Denver, CO, USA.10.1109/PHM.2008.4711422 ↗