Purdue University Graduate School
Browse
ThesisDocument_JFeng.pdf (12.96 MB)

VIBRO-ACOUSTIC MONITORING OF NESTED PLANETARY GEAR TRAINS

Download (12.96 MB)
thesis
posted on 2022-03-11, 13:42 authored by Jianxiong FengJianxiong Feng

Gear systems are major components in the modern power transmission system, especially in the automobile industry. In recent years, with the auto transmission function development, more compound gears have been utilized for better transmission speed ratio control and space-saving. The nested planetary gear train is widely applied among the newly designed compound gear trains due to its compactness. However, the nested gear train noise and vibration issue have not been studied extensively. In the current study, the vibrational and acoustical monitoring prototype will be developed to monitor the nested planetary manufacturing accuracy in the production line. Firstly, a novel testing fixture with the vertical and open setup is proposed to be used to monitor the gear, which is different from most gear condition monitoring studies. The open setup allows the accelerometer to be mounted directly on different clutches to monitor both carriers in the nested two-stage gear system closely. The test setup also enables the acoustical array to be implemented. Gear carriers with different pinion faults or damages are tested with both vibrational and acoustical sensors to identify the pinion gear error type and location. The vibrational data are processed with several classical rotating machinery signal processing techniques, including: (1) time synchronous averaging; (2) modulation sideband analysis; (3) modulation sideband modeling; (4) narrowband demodulation. Those techniques are also partially modified to adapt for the nested planetary structure monitoring. The results show that the unground or damaged pinions can be successfully identified and localized. Besides vibrational signal monitoring, an acoustical array is also proposed for noise source visualization and localization of the outer gearset (stage-2). The virtual rotating array (VRA) methods are proposed to de-Ddopperize the rotating source sound signals measured at stationary receivers. The VRA signals are subsequently used as input of the equivalent source based compressive sensing holography for high-resolution localization of the compact rotating sources. Both time-domain VRA and frequency domain VRA methods are studied and further developed for improved computational efficiency in terms of the VRA process. The new developments of the acoustical imaging system for rotating source localization are validated numerically and experimentally. Last but not least, the acoustical imaging system has also been successfully applied to the unground gear condition monitoring in the outer gearset of the nested gear train.

History

Degree Type

  • Doctor of Philosophy

Department

  • Mechanical Engineering

Campus location

  • West Lafayette

Advisor/Supervisor/Committee Chair

Kai Ming Li

Advisor/Supervisor/Committee co-chair

Yangfan Liu

Additional Committee Member 2

Stuart Bolton

Additional Committee Member 3

Fabio Semperlotti

Additional Committee Member 4

Robert Nowack