Application of Fast ICA and wavelet packet in ball mill vibration signal
Zhihong Jiang, Bo Hu · 6th International Workshop on Advanced Algorithms and Control Engineering (IWAACE 2022) · 2022
Aiming at the problem that the vibration signal of the ball mill is non-linear and non-stationary, which makes the load state difficult to identify, this paper proposes a method combining FastICA and wavelet packet to detect the load state of the ball mill. First, use the FastICA algorithm to denoise the collected original vibration signal; then use the demy wavelet basis function to perform wavelet packet energy spectrum analysis on the denoised vibration signal to obtain the vibration signal of each frequency band, and calculate each frequency band to the energy ratio of the internal vibration signal. By analyzing the vibration signals of the ball mill under different load conditions, the most sensitive characteristic frequency band with the change of the mill load is obtained. Finally, in the characteristic frequency band, the relationship model between energy and mill load is established to achieve the purpose of mill load detection. The experimental results show that the energy proportions of the characteristic frequency bands between the three load states of the ball machine are very different, which can well identify the load state of the ball mill and has practical application value.