8 April 2016 Gearbox fault diagnosis based on time-frequency domain synchronous averaging and feature extraction technique
Author Affiliations +
Abstract
Gearbox is one of the most vulnerable subsystems in wind turbines. Its healthy status significantly affects the efficiency and function of the entire system. Vibration based fault diagnosis methods are prevalently applied nowadays. However, vibration signals are always contaminated by noise that comes from data acquisition errors, structure geometric errors, operation errors, etc. As a result, it is difficult to identify potential gear failures directly from vibration signals, especially for the early stage faults. This paper utilizes synchronous averaging technique in time-frequency domain to remove the non-synchronous noise and enhance the fault related time-frequency features. The enhanced time-frequency information is further employed in gear fault classification and identification through feature extraction algorithms including Kernel Principal Component Analysis (KPCA), Multilinear Principal Component Analysis (MPCA), and Locally Linear Embedding (LLE). Results show that the LLE approach is the most effective to classify and identify different gear faults.
© (2016) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Shengli Zhang, Shengli Zhang, Jiong Tang, Jiong Tang, } "Gearbox fault diagnosis based on time-frequency domain synchronous averaging and feature extraction technique", Proc. SPIE 9804, Nondestructive Characterization and Monitoring of Advanced Materials, Aerospace, and Civil Infrastructure 2016, 98040K (8 April 2016); doi: 10.1117/12.2219460; https://doi.org/10.1117/12.2219460
PROCEEDINGS
9 PAGES


SHARE
Back to Top