Fault Signal Enhancement using Adaptive Filtering Techniques for Gearbox Monitoring

Vibration analysis that is the main condition monitoring techniques for machinery maintenance and fault diagnosis, this technique provides useful reliable information in machine condition monitoring program in practical application. In this paper, an adaptive algorithm is developed to preprocess the vibration signals of a gearbox operating under different conditions. Vibration signals gathered under different loads of the gearbox flange are considered as the input signals of the adaptive scheme, and the vibration signals with a certain time delay are used as the reference signal of the iterative adaptive scheme to attenuate the random and broadband noise in the vibration signals and hence to produce accurate detection results. The adaptive filter could update its coefficients adaptively to minimize the error signals subject to least
mean square (LMS) algorithm. The error signals include vital transient information of the vibration signals when the gearbox running under different conditions and the transient components can be enhanced after the adaptive process. It is a novel method to employ the residual signals for gearbox faults detection. Experimental data sets of gearbox vibration signals have been studied using adaptive filtering techniques. Preliminary results show that the adaptive filtering method is contribute to produce accurate results and hence obtain better detection and diagnosis results.