Generalized Exponential Degradation Model for a Tradeoff Between Prior and Posterior Model Parameters Distributions
IEEE Transactions on Instrumentation and Measurement · 2025 · 74 · 1-11
DOI: 10.1109/tim.2024.3509603 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Instrumentation and Measurement, 2025. Verfasst von Yikai Chen und 4 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 34 Literaturangaben. Gefördert von: National Key Research and Development Program of China, National Natural Science Foundation of China, Lenovo-Shanghai Jiao Tong University Strategic Cooperation Program.
Abstract (Originalsprache)
Exponential degradation modeling and its variants have been widely used to describe component or device degradation characteristics, such as exponential bearing degradation, and their extrapolations to a predefined failure threshold can be used to predict remaining useful life (RUL). Previously, a pioneered Bayesian exponential degradation model with random coefficients was developed to integrate online condition monitoring data with prior model parameters distributions to obtain posterior model parameters distributions so as to better describe an individual exponential component degradation process. In this article, a generalized exponential degradation model for a tradeoff between prior and posterior model parameters distributions is constructed to extend the previous pioneered exponential degradation model with random coefficients. Moreover, two new propositions are proposed to show that when hyperparameters used in the generalized exponential degradation model tend to 0, the posterior analytical parameters distributions of the generalized exponential degradation model are reduced to the posterior analytical parameters distributions of the previous pioneered exponential degradation model. When hyperparameters used in the generalized exponential degradation model tend to positive infinite, the posterior analytical parameters distributions of the generalized exponential degradation model are reduced to the prior model parameters distributions of the previous pioneered exponential degradation model. Overall, the proposed generalized exponential model can realize a tradeoff between prior and posterior model parameters distributions. Finally, the simulated and experimental cases are studied to illustrate how the proposed generalized exponential model works and demonstrate its effectiveness and superiority to exponential bearing degradation and RUL prediction.
Förderung
- National Key Research and Development Program of China
- National Natural Science Foundation of China
- Lenovo-Shanghai Jiao Tong University Strategic Cooperation Program
Zitieren
APA
Chen, Y., Wang, D., Liu, J., Huang, Q., & Lu, M. (2025). Generalized Exponential Degradation Model for a Tradeoff Between Prior and Posterior Model Parameters Distributions. IEEE Transactions on Instrumentation and Measurement, 74, 1-11. https://doi.org/10.1109/tim.2024.3509603
Vancouver
Chen Y, Wang D, Liu J, Huang Q, Lu M. Generalized Exponential Degradation Model for a Tradeoff Between Prior and Posterior Model Parameters Distributions. IEEE Transactions on Instrumentation and Measurement. 2025;74:1-11. doi:10.1109/tim.2024.3509603
BibTeX
@article{Chen2025,
author = {Chen, Yikai and Wang, Dong and Liu, Jie and Huang, Qingyun and Lu, Ming},
title = {Generalized Exponential Degradation Model for a Tradeoff Between Prior and Posterior Model Parameters Distributions},
journal = {IEEE Transactions on Instrumentation and Measurement},
year = {2025},
volume = {74},
pages = {1--11},
doi = {10.1109/tim.2024.3509603}
}
EUVIMED