Features extraction has always been crucial in rotary machines for Condition based machine health monitoring. Time-domain-segmentation being among the preliminary steps for further classification process plays a momentous role. Vibration signals from bearing are quasistationary in nature therefore calculation of constituent frequencies amplitudes in the vibration signal is dependent upon time-segmentation-window size. The proposed research confers the effects of time-segmentation window size on spectral features amplitudes calculation and its impacts on classification accuracy of the Artificial Neural Network (ANN). Using multi-size time-segmentation-window, for comprehensive spectral features calculation, ANN pattern classifier has been trained for enhanced classification. ANN learning assigns importance based relative weights to the links using supervised learning. Experimental results have shown that multi-size-window spectral features for ANN fault classifier perform efficiently for quasi-stationary bearing vibrations.
Vibration signals used for abnormality detection in machine health monitoring (MHM) are non-stationary in nature. Wavelet packet transform is extensively used in the literature for comprehensive analysis of non-stationary vibration signal but these techniques work only for a specific application lacking in some generalized methodology for selecting appropriate wavelet decomposition level and nodes for optimal performance. This study proposes a framework for inchoate fault detection by selecting the optimal wavelet decomposition level and nodes, named Optimally Parameterized Wavelet Packet Transform (OPWPT). OPWPT uses support vector machine to build the fault diagnostic model. Results in comparison with the existing schemes validate that OPWPT enhances the fault detection accuracy significantly in case of incipient faults when vibration signatures are very weak and overall signal to noise ratio is very poor.