Application on adaptive filtering and correlation filtering in impulse response signal feature extraction

Recognition of impulse response signals has important meaning to modal analysis and fault diagnosis of machinery. Impulse response signals are weak relatively and difficult to be measured because they are usually buried in the noise and signals of operation frequency and its harmonics with commensurate (or nearly commensurate) levels. Adaptive filter with parameters appropriately chosen is used to extract impulse response signal, which also acts as signal pre-processing. Correlation filtering based on Laplace wavelet is used to process the signal after adaptive filtering to extract vibration modal information from these signals. The processing results prove that the combination of adaptive filtering and correlation filtering based on Laplace wavelet can be effectively applied to extraction of impulse response signals.