EUVIMEDEUVIMED entdecken
Crossref

A RUL prediction method for rotating machinery based on adaptive TSP recognition and AE-IT-BiLSTM

Liyao Yu, Guofa Li, Jialong He, Jia Li

Measurement Science and Technology · 03.07.2026 · 37 (27) · 276106

DOI: 10.1088/1361-6501/ae804b ↗

Kurz gefasst

Fachartikel in der Zeitschrift Measurement Science and Technology, online erschienen am 06.07.2026. Verfasst von Liyao Yu und 3 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 37 Literaturangaben. Gefördert von: National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Rotating machinery has been widely used in modern industrial systems. Accurately predicting its remaining useful life (RUL) can not only reduce maintenance costs, but also improve operational safety. This study proposes a new adaptive prediction framework for RUL of rotating machinery. Firstly, dynamic time warping is used to calculate the minimum distance between the feature matrices before and after sensitive points, and adaptive detection of time to start prediction is achieved through a combination of empirical rules and data-driven learning. This hybrid strategy provides stronger adaptability and robustness across varying operating conditions without manual threshold adjustment. Subsequently, empirical mode decomposition is applied to separate low-frequency components and extract features that better reflect long-term degradation trends based on energy scales, effectively mitigating noise interference and improving feature reliability. Then, a multi-scale monotonic sample entropy index was developed to evaluate the contribution of each feature, achieving dynamic weighting of input features while avoiding the opaque and unreliable calculation process of traditional methods. Finally, these weighted features are fed into the proposed autoencoder and interactive bidirectional long short-term memory (AE-IT-BiLSTM) network to achieve accurate RUL prediction. Experimental results on both a public dataset and a private dataset demonstrate that the proposed framework significantly outperforms existing methods in prediction accuracy, confirming its effectiveness and strong generalization capability.

Förderung

  • National Natural Science Foundation of China

Zitieren

APA

Yu, L., Li, G., He, J., & Li, J. (2026). A RUL prediction method for rotating machinery based on adaptive TSP recognition and AE-IT-BiLSTM. Measurement Science and Technology, 37(27), 276106. https://doi.org/10.1088/1361-6501/ae804b

Vancouver

Yu L, Li G, He J, Li J. A RUL prediction method for rotating machinery based on adaptive TSP recognition and AE-IT-BiLSTM. Measurement Science and Technology. 2026;37(27):276106. doi:10.1088/1361-6501/ae804b

BibTeX
@article{Yu2026,
  author = {Yu, Liyao and Li, Guofa and He, Jialong and Li, Jia},
  title = {A RUL prediction method for rotating machinery based on adaptive TSP recognition and AE-IT-BiLSTM},
  journal = {Measurement Science and Technology},
  year = {2026},
  volume = {37},
  number = {27},
  pages = {276106},
  doi = {10.1088/1361-6501/ae804b}
}
Kostenlos starten →