TMS320 DSP based wavelet packet decompose on vibration signature analysis of turbine

The wavelet packet transformation (WPT) is a very powerful tool for the analysis of vibration signals. The WPT is introduced as an alternative means of extracting time-frequency information from vibration signals. The results of WPT coefficients provide us with arbitrary time-frequency resolution of a signal. However, WPT algorithm is very complicated, and the huge data computing needs lots of time. If we realize WPT with common CPU, it will need much time on computing, thus real-time vibration signals analysis become impossible. This paper illustrates a successful application of DSP based WPT algorithm. Some key issues on how to implement the complicated algorithm are thoroughly discussed, such as the realization of fast wavelet transformation, the optimization of the program structure. Extensive experiments show that the DSP based application meets the real-time requirement of vibration signals analysis, and It is a feasible and effective solution.