Orbit identification method based on ISOMAP for rotor system fault diagnosis

Orbit identification plays an important role for rotor system fault diagnosis. Due to signals noise and the nonlinear working condition of the rotor system, the orbit information which indicates the fault is very difficult to have got. A new method based on isometric feature mapping (ISOMAP) of orbit identification is provided in this paper. Firstly, the vibration signal of each orbit in a two-dimensional space is defined as a dimension feature of the high-dimensional space; the high-dimensional characteristics space is reduced into lower-dimensional space based on ISOMAP algorithm. Then the manifold features of the system orbit are obtained in order to identify the working conditions. The experimental results for a rotor test system proved the method designed in the paper effectively and validity. It provides a new method for fault diagnosis of rotor system.