Prediction of Bearing Wear in High-speed Aviation Fuel Pump Using a VMD-CNN-LSTM Model
Journal of Applied Fluid Mechanics · 01.09.2026 · 19 (9)
DOI: 10.47176/jafm.19.9.4268 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Applied Fluid Mechanics, online erschienen am 01.09.2026. Verfasst von X. Wang und 4 weiteren Autor:innen.
Abstract (Originalsprache)
The high-speed aviation fuel pump is a crucial component of the aircraft fuel system, and the condition of its sliding bearings has a direct impact on the reliability of aero-engines and the overall flight safety. This study presents a novel predictive framework designed to accurately monitor and forecast the end-face wear of sliding bearings in high-speed aviation fuel pumps. This framework integrates variational modal decomposition (VMD), convolutional neural networks (CNN), and long short-term memory (LSTM) networks into a combined model named VMD-CNN-LSTM. VMD adaptively decomposes the original wear signal into six intrinsic mode functions, aiding in the reduction of measurement noise and enhancing signal stability. CNN obtains localized degradation features from these IMFs, while LSTM identifies long-term temporal patterns in wear progression. This results in a thorough approach that includes signal decomposition, feature extraction, and time-series prediction. The proposed model was compared with LSTM and VMD-LSTM models on five complete lifecycle wear datasets from high-speed aviation fuel pumps. VMD-CNN-LSTM demonstrated enhanced performance: root mean square error of 0.4×10-3 and mean absolute percentage error of 0.29%, representing improvements of 81.8% and 84.7%, respectively, over the conventional LSTM model. Thus, the integration of noise-resistant decomposition of VMD with the ability of CNN to learn spatial features and dynamic sequence modeling of LSTM significantly enhances the precision and dependability of sliding-bearing wear forecasts.
Zitieren
APA
Wang, X., Liu, W., Ma, P., Chai, L., & Guo, G. (2026). Prediction of Bearing Wear in High-speed Aviation Fuel Pump Using a VMD-CNN-LSTM Model. Journal of Applied Fluid Mechanics, 19(9). https://doi.org/10.47176/jafm.19.9.4268
Vancouver
Wang X, Liu W, Ma P, Chai L, Guo G. Prediction of Bearing Wear in High-speed Aviation Fuel Pump Using a VMD-CNN-LSTM Model. Journal of Applied Fluid Mechanics. 2026;19(9). doi:10.47176/jafm.19.9.4268
BibTeX
@article{Wang2026,
author = {Wang, X. and Liu, W. and Ma, P. and Chai, L. and Guo, G.},
title = {Prediction of Bearing Wear in High-speed Aviation Fuel Pump Using a VMD-CNN-LSTM Model},
journal = {Journal of Applied Fluid Mechanics},
year = {2026},
volume = {19},
number = {9},
doi = {10.47176/jafm.19.9.4268}
}
EUVIMED