Parametric Bispectrum Analysis of Cracked Rotor Based on Blind Identification of Time Series Models

The crack is a very serious defect in the rotor system. However, compared with any other defects, diagnosis of the rotor crack is more difficult, and its potential harm is similarly very serious. Therefore, it is very urgent and necessary to detect and diagnose the rotor crack effectively. In this paper, an approach of bispectrum analysis based on blind identification of time series model is proposed, and this approach is then applied to the fault diagnosis of the rotor crack. The bispectrum characteristics of the cracked rotor with cracks of different depth and location are investigated, and some valuable results are obtained. This approach provides a new idea and method for the diagnosis of the rotor crack. The experimental results show that this approach is very effective