Condition health monitoring of dynamic Systems based on vibration signatures has generally relied upon Fourier based analysis as a means of translating vibration signals in time domain into the frequency domain. However, Fourier analysis provideda poor representation of signals well localized in time. In this case, it is difficult to detect and identify the signal pattern from their expansion coefficients because the information is diluted across the whole basis. The wavelet packet transform isintroduced as an alternative means of extracting time-frequency information from vibration signatures. Moreover, with the aid of statistical based feature selection criteria, a lot of feature components containing little discriminant information could bediscarded resulting in a feature subset with reduced number of parameters. This significantly reduces the long training time that is often associated with neural network classifier and increases the generalization ability of the neural network classifier. To validate the feature extraction algorithm proposed, the simulations have been performed with the benchmark data known as Westland vibration data set. The results show significant improvement when the data is subjected to various white, colored and pink noise.
展开▼
