Research on fault diagnosis of gearbox under noise environment based on particle filter

In the gearbox fault diagnosis, the signal collected by sensors is generally suffered by the disturbance from various types of unknown noise. Under the complex noise environment, the blind source separation of the gearbox fault cannot obtain perfect result of separation. In order to solve this problem, a new noisy blind source separation method of gearbox fault based on particle filter is proposed. A denoising processing to the observation signal is implemented using Rao-blackwellised particle filter before the independent component analysis. The simulation and experimental results show that the proposed method is effective.