Analysis of Vibration Signal Fault Modes based on Improved Deep Belief Network

To improve the calculation accuracy and speed of aero-engine rotor vibration fault identification, an improved DBN fault modes recognition method based on FPCD and Gaussian convolution RBM is proposed. A small sample identification model based on DBN is established to identify rotor vibration fault modes. In this paper, eigenvalues are extracted by using FFT and waveWe let packet transform first. Then, the multi-domain features are taken as input items to train and recognize using improved DBN. Finally, an algorithm model that can effectively identify the fault mode of the rotor vibration signal is established. The results show that the improved DBN in this paper can effectively classify and recognize different vibration signal patterns, and the correctness rate is 97.5%.