In this paper, theoretical foundations of fractal compression and de-noise are introduced. Fractal property of the vibration signals is analyzed. An adaptive fractal de-noise algorithm is presented based on fuzzy filtering theory and fractal dimension. All of the fractal coding parameters are estimated. An adaptive algorithm for vibration signal data compression is proposed based on Iterated Function Systems (IFS) and College theorem. The peculiarity of the adaptive fractal algorithm is that the length of the subsection decided by error threshold in terms of property of vibration signals. This method is applied on simulation signals, experimental data and on-site data. It shows that the algorithm promoted in this paper is practicable according to analyzing applications of these vibration signals, computing compression ratios and signal-to-noise of reconstructed signal. All of parameters meet the follow-up requirements of fault diagnosis. |