Adaptive tacho information estimation for surveillance of rotatory machine under nonstationary conditions

International audienceRolling bearing faults are the leading causes of downtime in rotary machines. In recent years, numerous and various vibration-based approaches have been put forwarded for rolling bearing fault detection. In the vibration-based techniques, order tracking-based methods are considered as very effective techniques. In the current reported order tracking methods, auxiliary devices are essential to obtain the instantaneous angular speed (IAS) of the machine. Aiming at this shortcoming, estimating IAS from vibration signals has been studied and some tacho-less order tracking (TLOT) techniques have been put forwarded. However, the effectiveness of the current available TLOT algorithms rely on the manually selection of the initial parameters for IAS estimation, which bring about user-friendliness. In order to tackle the aforementioned obstacles, a novel adaptive tacho information estimation method based on nonlinear mode decomposition (NMD) is proposed. In the proposed method, the nonlinear mode decomposition (NMD) method is improved and its computational burden is reduced. And then, the tacho information is adaptively estimated. The vibration signal collected from an aircraft engine is used for signal analysis and the effectiveness of the proposed is successfully validated