Quantcast

Application of the state deterioration evolution based on bi-spectrum entropy and HMM in wind turbine

Research paper by Xiuli Liu, Xiaoli Xu, Zhanglei Jiang, Guoxin Wu, Yunbo Zuo

Indexed on: 14 Mar '16Published on: 14 Nov '15Published in: Chaos, Solitons & Fractals



Abstract

Concerning the problem of large rotating machinery with non-stationary state like wind turbine, this research mainly makes an emphasis on the method of state deterioration recognition based on bi-spectrum entropy and HMM (Hidden Markov Model). Firstly, the true signal such as low-speed start vibration signals of rotor test rig in the normal state and a plurality of imbalance deterioration degrees are collected. Bi-spectrum is applied to obtain the fault feature from the vibration signals mixed with a complex background noise. On the basis of bi-spectrum analysis, a bi-spectrum entropy algorithm is derived under the condition of subspace distribution probability, and the HMM for the fault pattern recognition is established by using the bi-spectrum entropy feature as input. This method is verified by successfully recognizing four state deterioration degrees. Finally, the method is applied to recognize the imbalance deterioration degree of wind turbine with the type of SL1500/82 and equipment actual working condition verified the effectiveness of the proposed method.

Figure 10.1016/j.chaos.2015.10.018.0.jpg
Figure 10.1016/j.chaos.2015.10.018.1.jpg
Figure 10.1016/j.chaos.2015.10.018.2.jpg
Figure 10.1016/j.chaos.2015.10.018.3.jpg
Figure 10.1016/j.chaos.2015.10.018.4.jpg
Figure 10.1016/j.chaos.2015.10.018.5.jpg
Figure 10.1016/j.chaos.2015.10.018.6.jpg
Figure 10.1016/j.chaos.2015.10.018.7.jpg