Application of wavelet fractal algorithm to feature extraction of hydro-turbine vibration signals

作者：Zhihuai X^*; Yufan C; Yanan Y; Malik O P

来源：Journal of Optoelectronics and Advanced Materials, 2015, 17(1-2): 93-102.

• Fractal dimension
• Correlation dimension
• Wavelet fractal algorithm
• Hydropower vibration signals

To avoid the drawbacks of traditional fractal theory, such as complicated calculation and the difficulty in choosing the proper kind of fractal dimension, a wavelet fractal algorithm is proposed in this paper for the feature extraction of a hydro-turbine vibration signals. In this algorithm, wavelet functions are used to decompose the de-noised signal. After decomposition, variance of each level of the detailed components is introduced to describe the energy distribution on each level. The fractal dimension is the slope of the fitting line by taking scale j as the horizontal axis and variance as the vertical axis. To verify the theory introduced in this paper, a comparison of the wavelet fractal algorithm with the conventional fractal algorithm on a few sets of experimental vibration signals shows that although both methods are successful in feature extraction, the wavelet fractal algorithm provides more accurate feature extraction of hydro-turbine vibration signals.

• 出版日期2015-2
• 单位武汉大学