A Novel Trajectory Prediction Approach for the Active Magnetorheological Fluid Bearing-Rotor System based on VMD-IGWO-LSTM
ID:126 Submission ID:131 View Protection:ATTENDEE Updated Time:2023-06-01 11:22:46 Hits:518 Oral Presentation

Start Time:2023-06-10 14:40 (Asia/Shanghai)

Duration:15min

Session:[S1] Concurrent Session 1 » [S1-5] Concurrent Session 1-5

No files

Abstract
In order to address the problems of insufficient load capacity and rotor vibration of large grinding ball mill, an active fluid-film bearing lubricated with magnetorheological fluid (MRF) is proposed. Firstly, the geometry of the MRF bearing is designed and its intelligent lubrication mechanism is analyzed to clarify its advantages. In addition, mathematical model of MRF fluid-film bearing-rotor system is derived to select the appropriate variable parameters as inputs and outputs of training model, and the FEM simulation is utilized to obtain the dataset of rotor trajectory in COMSOL Multiphysics. Moreover, a novel prediction approach based on variational mode decomposition (VMD), improved grey wolf optimization (IGWO) and long short-term memory (LSTM), namely VMD-IGWO-LSTM, is proposed to predict the rotor trajectory of the active MRF bearing-rotor system in this work. Finally, the experiments demonstrate the effectiveness of the proposed method compared with other methods.
 
Keywords
Magnetorheological fluid,fluid-film bearing,variational mode decomposition,improved gray wolf optimization,Long short-term memory
Speaker
Peng Lai
student China University of Mining and Technology

Submission Author
Peng Lai China University of Mining and Technology
Shen Yurui China University of Mining and Technology
Wang Qiyu China University of Mining and Technology
Hua Dezheng China University of Mining and Technology
Liu Xinhua China University of Mining and Technology
Comment submit
Verification code Change another
All comments