Quantcast

Trends in non-stationary signal processing techniques applied to vibration analysis of wind turbine drive train – A contemporary survey

Research paper by R. Uma Maheswari, R. Umamaheswari

Indexed on: 18 Oct '16Published on: 24 Aug '16Published in: Mechanical Systems and Signal Processing



Abstract

Publication date: 15 February 2017 Source:Mechanical Systems and Signal Processing, Volume 85 Author(s): R. Uma Maheswari, R. Umamaheswari Condition Monitoring System (CMS) substantiates potential economic benefits and enables prognostic maintenance in wind turbine-generator failure prevention. Vibration Monitoring and Analysis is a powerful tool in drive train CMS, which enables the early detection of impending failure/damage. In variable speed drives such as wind turbine-generator drive trains, the vibration signal acquired is of non-stationary and non-linear. The traditional stationary signal processing techniques are inefficient to diagnose the machine faults in time varying conditions. The current research trend in CMS for drive-train focuses on developing/improving non-linear, non-stationary feature extraction and fault classification algorithms to improve fault detection/prediction sensitivity and selectivity and thereby reducing the misdetection and false alarm rates. In literature, review of stationary signal processing algorithms employed in vibration analysis is done at great extent. In this paper, an attempt is made to review the recent research advances in non-linear non-stationary signal processing algorithms particularly suited for variable speed wind turbines.