Research on generating unit vibration signal by means of wavelet-fractal theory

Wavelet or wavelet packet transform and fractal theory are consistent in nature and in process to understand things, so it is feasible to blend both of them felicitously to create a new technology wavelet-fractal. In this paper, wavelet-fractal is used to analyze and study malfunction of generating unit rotor synthetically. Given three time domain vibration signals of one 28 MW generating unit at different running state, decomposing, reconstructing and calculating the correlation dimensions of these signals, some useful information can be found. The research result shows that the compound of wavelet packet decomposition and reconstruction with fractal technique is powerful to understand the running state of the generating unit clearly. So this method can be used as synthesis quantify index to depict mechanical fault characteristic of generator set is educed. The research can be generalized to any nonlinear dynamical system and gives a new approach for monitoring condition, fault diagnosis and prediction online.