Abstract
We developed a fatigue residual useful life (RUL) prediction model using the available time-series fatigue data of Ni-base alloy welds via a long short-term memory (LSTM) network. The effects of some LSTM network hyperparameters on model performance were investigated through sensitivity studies. The LSTM network model outperformed multiple regression models when the LSTM model hyperparameters were appropriately tuned. However, the additional gain was insignificant, considering that the LSTM network was much more complex than multiple regression models. The best performance of the LSTM network model was achieved when the number of hidden units, input window size, and batch size were small and the number of LSTM layers was large.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
M. Benson, D. Rudland and A. Csontos, Weld Residual Stress Finite Element Analysis Validation: Part 1 - Data Development Effort (NUREG-2162), U.S.NRC, Washington DC (2014).
T. Maeguchi et al., PWSCC susceptibility of Alloy 690, 52 and 152, Proceedings of the 18th International Conference on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactors, Springer (2019) 485–500.
S. Mohanty et al., Tensile behavior of 82/182 filler, butter and heat-affected-zones in a 508 LAS-316 SS dissimilar weld: tensile test, Material Model and Finite Element Model Validation, in Pressure Vessels and Piping Conference, American Society of Mechanical Engineers (2019) V001T01A024.
ASME, Rules for Construction on Nuclear Facility Components: Supports III Division 1-Subsection NF, American Society of Mechanical Engineers (2013).
O. K. Chopra and G. L. Stevens, Effect of LWR Water Environments on the Fatigue Life of Reactor Materials (NUREG/CR-6909, Rev 1), United States Nuclear Regulatory Commission, Office of Nuclear Regulatory (2018).
J. P. Park et al., Statistical analysis of S-N type environmental fatigue data of Ni-base alloy welds using weibull distribution, Nuclear Engineering and Technology (2022) (In Press).
J. P. Park et al., Weibull and bootstrap-based data-analytics framework for fatigue life prognosis of the pressurized water nuclear reactor component under harsh reactor coolant environment, J. Nondestruct Eval Diagn Progn Eng. Syst, 3 (1) (2020).
A. Akrim et al., A framework for generating large data sets for fatigue damage prognostic problems, 2022 IEEE International Conference on Prognostics and Health Management (ICPHM) (2022) 25–33.
E. Zio, Diagnostics and prognostics of engineering systems: methods and techniques, Diagnostics and Prognostics of Engineering Systems: Methods and Techniques, IGI Global (2012) 333–356.
K. L. Tsui et al., Prognostics and health management: A review on data driven approaches, Math Probl Eng., 2015 (2015).
X.-S. Si et al., Remaining useful life estimation-a review on the statistical data driven approaches, Eur J. Oper Res, 213 (1) (2011) 1–14.
L. He et al., Machine learning-based predictions of fatigue life and fatigue limit for steels, J. Mater. Sci. Technol., 90 (2021) 9–19.
P. Shi, L. Hong and D. He, Using long short term memory based approaches for carbon steel fatigue remaining useful life prediction, 2018 Prognostics and System Health Management Conference (PHM-Chongqing), IEEE (2018) 1055–1060.
J. P. Park, S. Mohanty and C.B. Bahn, Cyclic rambergosgood parameters of 316SS weld metal under in-air and PWR-water environments, J. Press Vessel Technol., 143 (5) (2021).
S. Hochreiter and J. Schmidhuber, Long short-term memory, Neural Comput, 9 (8) (1997) 1735–1780.
S.-N. Nguyen et al., Neural network-based prediction of the long-term time-dependent mechanical behavior of laminated composite plates with arbitrary hygrothermal effects, Journal of Mechanical Science and Technology, 35 (10) (2021) 4643–4654.
M. K. Park et al., Predictive model for PV power generation using RNN (LSTM), Journal of Mechanical Science and Technology, 35 (2) (2021) 795–803.
U. E. Akpudo and J.-W. Hur, A feature fusion-based prognostics approach for rolling element bearings, Journal of Mechanical Science and Technology, 34 (10) (2020) 4025–4035.
T. Xia et al., An ensemble framework based on convolutional bi-directional LSTM with multiple time windows for remaining useful life estimation, Comput. Ind., 115 (2020) 103182.
J. Xia et al., LSTM-based multi-layer self-attention method for remaining useful life estimation of mechanical systems, Eng. Fail Anal., 125 (2021) 105385.
Y. Wu et al., Remaining useful life estimation of engineered systems using vanilla LSTM neural networks, Neurocomputing, 275 (2018) 167–179.
Y. Yu et al., Averaged Bi-LSTM networks for RUL prognostics with non-life-cycle labeled dataset, Neurocomputing, 402 (2020) 134–147.
H. Miao et al., Joint learning of degradation assessment and RUL prediction for aeroengines via dual-task deep LSTM networks, IEEE Trans. Industr. Inform., 15 (9) (2019) 5023–5032.
S. Hochreiter et al., Gradient flow in recurrent nets: the difficulty of learning long-term dependencies, S. C. Kremer and J. F. Kolen (eds.), A Field Guide to Dynamical Recurrent Neural-Networks, IEEE Press (2001).
J. Liu et al., Remaining useful life prediction of PEMFC based on long short-term memory recurrent neural networks, Int. J. Hydrogen Energy, 44 (11) (2019) 5470–5480.
H. Hu, H. Luo and X. Deng, Health monitoring of automotive suspensions: a lstm network approach, Shock and Vibration, 2021 (2021).
J. Ham et al., Low cycle fatigue life of alloy 52M weld metal in simulated PWR environment, Transactions of the Korean Nuclear Society Virtual Spring Meeting, Korean Nuclear Society (2021).
ASTM International, Standard Practice for Strain-Controlled Fatigue Testing, ASTM E606-04, ASTM International (2017).
S. Mohanty et al., Tensile and Fatigue Testing and Material Hardening Model Development for 508 LAS Base Metal and 316 SS Similar Metal Weld Under In-air and PWR Primary Loop Water Conditions (ANL/LWRS-15/02), Argonne National Lab. (ANL), Argonne, IL (United States), IL, USA (2015).
J.-L. Chaboche, Time-independent constitutive theories for cyclic plasticity, Int. J. Plast, 2 (2) (1986) 149–188.
B. Barua et al., A cyclic-plasticity-based mechanistic approach for fatigue evaluation of 316 stainless steel under arbitrary loading, J. Press Vessel Technol, 140 (1) (2018).
S. Brooks et al., Handbook of Markov Chain Monte Carlo, CRC Press, USA (2011).
W. F. Wu and C. C. Ni, Probabilistic models of fatigue crack propagation and their experimental verification, Probabilistic Engineering Mechanics, 19 (3) (2004) 247–257.
H. Hewamalage, C. Bergmeir and K. Bandara, Recurrent neural networks for time series forecasting: current status and future directions, Int. J. Forecast, 37 (1) (2021) 388–427.
MathWorks, IstmLayer, https://kr.mathworks.com/help/deeplearning/ref/nnet.cnn.layer.lstmlayer.html (2021).
E. Muccino, Input Window Size for Deep Recurrent Reinforcement Learning, Mindboard, https://Medium.Com/Mindboard/Input-Window-Size-for-Deep-Recurrent-Reinforcement-Learning-4e5d2960756b (2019).
O. Kuss, Global goodness-of-fit tests in logistic regression with sparse data, Stat Med, 21 (24) (2002) 3789–3801.
MathWorks, Filtering and Smoothing Data, https://kr.mathworks.com/help/curvefit/smoothing-data.html?lang=en (2022).
H. Wang et al., Remaining useful life prediction techniques for electric valves based on convolution auto encoder and long short term memory, ISA Trans., 108 (2021) 333–342.
N. Srivastava et al., Dropout: a simple way to prevent neural networks from overfitting, The Journal of Machine Learning Research, 15 (1) (2014) 1929–1958.
A. Krizhevsky, I. Sutskever and G. E. Hinton, Imagenet classification with deep convolutional neural networks, Adv. Neural Inf. Process Syst., 25 (2012) 1097–1105.
Acknowledgments
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT: Ministry of Science and ICT) (No. 2019M2A8A1000 64013), the Korea Institute of Energy Technology Evaluation and Planning (KETEP), and the Ministry of Trade, Industry & Energy (MOTIE) of the Republic of Korea (No. 20214000000410).
Author information
Authors and Affiliations
Corresponding author
Additional information
Jae Phil Park is a postdoctoral student at the School of Mechanical Engineering, Pusan National University, Busan, Korea. He received his Ph.D. in Mechanical Engineering from Pusan National University. His research interests include degradation of nuclear materials, safety evaluation, and life prediction using statistical methods.
Junhyuk Ham is a graduate student in the Department of Nuclear Engineering, Ulsan National Institute of Science and Technology, Ulsan, Korea. His research interests include corrosion fatigue of structural materials and irradiation-assisted stress corrosion cracking behavior of additively manufactured materials.
Ji Hyun Kim is a Professor at the Department of Nuclear Engineering, Ulsan National Institute of Science and Technology, Ulsan, Korea. He received his Ph.D. in Nuclear Engineering from Seoul National University. His research interests include commercial and next-generation reactor materials, and spent-nuclear-fuel pool materials.
Young Jin Oh is a senior researcher at the Smart Convergence Research Institute, KEPCO E&C Co., Ltd. He received his Ph.D. in Nuclear Engineering from Seoul National University. His research interests include structural integrity assessment and deep-learning applications to power plant engineering.
Chi Bum Bahn is an Associate Professor at the School of Mechanical Engineering, Pusan National University, Busan, Korea. He received his Ph.D. in Nuclear Engineering from Seoul National University. His research interests include material degradation and structural integrity of nuclear power systems.
Rights and permissions
About this article
Cite this article
Park, J.P., Ham, J., Kim, J.H. et al. Fatigue residual useful life estimation of Ni-base alloy weld with time-series data. J Mech Sci Technol 37, 2353–2362 (2023). https://doi.org/10.1007/s12206-023-0412-y
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12206-023-0412-y