The Study on Rotating Machinery Fault Diagnosis Based on Deep Neural Networks

In the paper, the author mainly applies vibration signals of whirl machinery rotors to do spectral analysis and fault diagnosis, selects 8 spectral values of characteristic frequent sections as the eigenvectors of diagnosis, and applies deep neural networks to analyze eigenvectors of vibration, so as to improve precision of judging deep neutral networks and make neutral networks have rules to follow in weight and threshold selection.